symptoms list | confidence float64 | log_events list | causal_chain list | incident_summary string | cause_vs_symptoms dict | root_cause_service string | root_cause_summary string | diagnostic_rationale string |
|---|---|---|---|---|---|---|---|---|
[
"Access tokens rejected as not active by the API gateway.",
"Sign-in requests failing with HTTP 401."
] | 0.94 | [
{
"timestamp": "2025-04-22T11:29:42.105Z",
"service_name": "identity-service",
"request_id": "req-a912",
"trace_id": "tr-d071",
"log_level": "WARN",
"message": "Token issuer clock offset is +18000 ms relative to the configured time source."
},
{
"timestamp": "2025-04-22T11:29:42.110Z... | [
{
"step_number": 1,
"service_name": "identity-service",
"event_description": "Identity-service reported its issuer clock was 18000 ms ahead of the configured time source.",
"causal_link": "The logged clock offset is consistent with the token's not-before time being later than the gateway's current t... | Sign-in requests were rejected with HTTP 401 because the gateway received access tokens whose not-before time was ahead of the gateway's current time. | {
"underlying_cause": "Identity-service reported a +18000 ms issuer clock offset and issued a token whose not-before time was ahead of the gateway's current time.",
"symptom_explanation": "The gateway's token rejection and the resulting HTTP 401 are effects of the time discrepancy; they are not independent causes o... | identity-service | Identity-service reported a clock offset of +18000 ms and issued a token with a not-before time of 11:30:00Z, while the gateway evaluated it at 11:29:42Z. The token was therefore not yet active according to the gateway's timestamp. | For the same request and trace, identity-service first logs a +18000 ms clock offset and then issues a token with nbf=11:30:00Z. The gateway subsequently compares that value with its own 11:29:42Z time and explicitly rejects the token as not active. Web-session-service then records the resulting HTTP 401. This direct e... |
[
"A growing backlog of unacknowledged invoice events.",
"Invoice projection data lagging behind the event stream.",
"Billing API invoice summary requests timing out."
] | 0.95 | [
{
"timestamp": "2025-03-06T16:42:10.012Z",
"service_name": "invoice-writer",
"request_id": "req-b204",
"trace_id": "tr-6c8e",
"log_level": "INFO",
"message": "Published invoice.updated event for invoice inv-5831 with tax_total=18.75."
},
{
"timestamp": "2025-03-06T16:42:10.038Z",
... | [
{
"step_number": 1,
"service_name": "invoice-writer",
"event_description": "Invoice-writer published an invoice.updated event with tax_total=18.75.",
"causal_link": "The decimal value did not match the integer type expected by the projection worker's decoder."
},
{
"step_number": 2,
"ser... | Invoice summaries became stale and billing API requests timed out after the invoice projection worker repeatedly rejected events whose decimal tax_total values did not match its expected integer type. | {
"underlying_cause": "A schema type mismatch in invoice-projection-worker: its decoder expected integer tax_total values, but the published event contained the decimal value 18.75.",
"symptom_explanation": "The backlog, stale invoice summaries, and billing API timeouts are downstream effects of the worker repeated... | invoice-projection-worker | The worker's event decoder expected tax_total to be an integer, while invoice-writer published a decimal value. The worker rejected the event and retried it without acknowledgment, preventing the consumer group from advancing and contributing to the growing backlog. | The invoice-writer log shows the decimal tax_total value, and the worker log explicitly identifies that value as incompatible with its expected integer type and says the message will be retried without acknowledgment. The event-bus backlog, stale projection, and billing API timeout occur later and fit the resulting pro... |
[
"Inventory lookup timeouts while waiting for a database connection.",
"Catalog availability lookups failing because the inventory dependency did not respond.",
"Product detail requests returning HTTP 503."
] | 0.96 | [
{
"timestamp": "2025-02-14T09:14:02.100Z",
"service_name": "inventory-service",
"request_id": "req-7f31",
"trace_id": "tr-91ac",
"log_level": "ERROR",
"message": "Database pool exhausted: 32 of 32 connections checked out; 27 leases exceeded the 300-second checkout limit."
},
{
"times... | [
{
"step_number": 1,
"service_name": "inventory-service",
"event_description": "The service logged that all 32 database connections were checked out and 27 leases had exceeded the 300-second limit.",
"causal_link": "With no connections available, subsequent inventory lookups could not acquire one and... | Product detail requests returned HTTP 503 after inventory lookups began timing out while the inventory database connection pool was fully occupied. | {
"underlying_cause": "Inventory-service had exhausted its database connection pool, with 27 of 32 checked-out leases exceeding the logged checkout limit.",
"symptom_explanation": "The timeouts and HTTP 503 responses occurred downstream after inventory-service could not provide data; they describe the impact of the... | inventory-service | The inventory database connection pool was exhausted: all 32 connections were checked out, and 27 had exceeded the recorded checkout limit. New inventory lookups consequently timed out waiting for a connection. | The earliest relevant event is the inventory-service error reporting a fully occupied connection pool and overdue leases. The same trace then shows an inventory lookup timing out, catalog-service reporting its inventory dependency timeout, and storefront-service returning HTTP 503. This sequence supports the pool exhau... |
Multi-Service Application Failure Root Cause Analysis Log Dataset
This dataset contains timestamped log events from multiple services, request-correlation information, and annotations including incident summaries, root cause services, underlying faults, cross-service causal chains, observed symptoms, and diagnostic evidence. Its structured labels distinguish underlying causes from visible symptoms, enabling models to reconstruct how failures propagate through distributed applications, identify root causes, and provide evidence-based explanations. It is suited to application performance management, service reliability analysis, and automated incident diagnosis.
Technical Specifications
| Field | Type | Description |
|---|---|---|
| symptoms | array | Observable errors or performance effects reported in the logs, distinct from the underlying cause. |
| confidence | number | Confidence in the root cause assessment, represented as a value from zero to one. |
| log_events | array | Timestamped log entries collected from the services involved in the incident. |
| causal_chain | array | Ordered sequence of events showing how the root cause propagated across services. |
| incident_summary | string | Concise account of the incident inferred from the raw log events. |
| cause_vs_symptoms | object | Explicit distinction between the underlying cause and the observable symptoms of the incident. |
| root_cause_service | string | Service identified as the incident's origin based on the available evidence. |
| root_cause_summary | string | Explanation of the underlying fault that initiated the incident. |
| diagnostic_rationale | string | Evidence-based reasoning connecting the log events to the root cause and causal chain. |
Compliance Statement
| Authorization Type | CC-BY-NC-SA 4.0 (Attribution–NonCommercial–ShareAlike) |
| Commercial Use | Requires exclusive subscription or authorization contract (monthly or per-invocation charging) |
| Privacy and Anonymization | No PII, no real company names, simulated scenarios follow industry standards |
| Compliance System | Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs |
Source & Contact
- Downloads last month
- -