text
stringclasses 4
values |
|---|
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |
Snappyflow is an application performance monitoring tool |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Snappyflow provides us the ability to 1. monitor metrics like CPU, RAM utilization, etc. 2. collect application logs 3. perform application tracing 4. perform profiling |
Heirarchy of objects in Snappyflow is as follows: 1. Each customer is given a profile. A profile is a combination of Kafka topics, an S3 bucket and a set of indices. 2. A profile can consist of one or more projects. As the name suggests, a project inidcates a set of applications like database, infrastructure the project runs on etc. 3. Each project consists of one or more application. An application can be thought of as a set of entities which can be monitored atomically e.g., frontend, nginx + Elasticsearch upstream, Kubernetes pods belonging to a specific tier of your project like Archival etc. 4. Every application, will consist of a dashboard to visualize metrics, view the logs, view the service mesh of that application along with inter-communication latency figures. 5. Under the logs, you can manage datasets. A dataset is a combination of plugin and document-type e.g., elasticsearc-logs is a plugin and one of the document-types is GC-Logs A dataset can go to Primary Storage or/and Secondary Storage Primary storage uses Elasticsearch. Hence, we get a very low search latency. Primary storage is much more expensive than secondary storage. Secondary storage uses cloud Blob storage. Hence, we get a higher search latency than Primary storage |
Internals of snappyflow are as follows: 1. When we create a profile in Snappyflow, it internally creates a set of kafka topics. One for each data type e.g., log topic to handle logs of whole profile (all projects under this profile share the topic), metric topic to handle metrics, etc,. It also creates a kafka-connect elasticsearch connector to poll data from Kafka and write it into primary storage. 2. When we create a project under a profile in Snappyflow, it internally creates a set of indices. One for each data type e.g., log index to handle logs (all applications under this project share the index) It alters the primary-storage kafka-connect connector to mark this project as a whitelist 3. Under an application, if we create a rule to store dataset-1 in primary as well as secondary storage, it creates another kafka-connect connector to poll data from Kafka and write it into secondary storage. 4. Snappyflow consists of the following modules Frontend -> this is the graphical-user-interface. It is deployed as a pod in Kubernetes under apm namespace Platform WebApp -> this is a pod deployed in Kubernetes under apm namespace. It helps to perform CRUD operations on profiles, projects, applications and handles user RBAC Platform Celery -> this consists of several celery pods which are listening on multiple celery queues. it helps to perform async tasks like auto-discovery of project/application, Alert count accumulation, etc,. VizBuilder -> this is a set of pods which are present under apm namespace in Kubernetes. this module helps to convert snappyflow query DSL into elasticsearch query DSL for querying from primary storage. It does a similar operation for secondary storage. it converts snappyflow DSL into SQL for use in Presto. Elasticsearch -> this serves as a primary store for Snappyflow Presto -> this serves as an SQL engine to search secondary storage. Secondary storage is nothing but cloud blob storage. Datapath -> This consists of Zookeeper, Kafka, Kafka-rest-proxy which is a HTTP interface which is used by snappyflow agents to write data into kafka, 2 Kafka-Connect clusters (one for Primary storage and the other for secondary storage) and a kafka-controller which is a custom application to manage this kafka ecosystem via CRUD APIs Archival -> this is a set of pods which detect the data in blob/secondary storage and perform necessary operations to make it searchable. We write data into secondary storage in a time based directory format i.e., dataset-1/year=2023/month=10/day=20 etc. Archival pods detect these folders and add these partitions into hive server. Once the partitions are added to hive, presto will be able to search this data |