train
stringlengths
20
1.23k
"Shaun Keenan":"on my 3080"
"Shaun Keenan":"It uses the copilot API"
"Shaun Keenan":"with Envoy? wow I hadn’t thought of that. very nice"
"Shaun Keenan":"Leaving in 30 mins for baby ultrasound and tests. Long appointment so I will be out for a couple hours"
"Shaun Keenan":"the answer, I’m guessing, is no :stuck_out_tongue: - is there a way to get project quotas set up immediately after the project is created, and to do so in code?"
"Shaun Keenan":"got a customer with a big stack that they want to deploy in one shot (very repeatably). will probably pre-create projects in advance if not"
"Shaun Keenan":"not API quotas but project quotas, e.g. N1 cores & RAM"
"Shaun Keenan":"also these limit existing quotas:\n> A consumer override is applied to the consumer on its own authority to limit its own quota usage"
"Shaun Keenan":"I remember looking into this for Sauce, it’s really annoying that they don’t have any resources/API’s for this"
"Shaun Keenan":"<https://www.willett.io/posts/precepts/>"
"Shaun Keenan":"yeah, you can use `--flatten` to show more"
"Shaun Keenan":"one sec"
"Shaun Keenan":"sample, could add a `--filter` on the owner role to this\n```$ gcloud projects get-iam-policy sk-sandbox01 --flatten=\"bindings[].members\" --format='table(bindings.role,bindings.members)'\nROLE MEMBERS\nroles/artifactregistry.serviceAgent serviceAccount:service-874864430437@gcp-sa-artifactregistry.iam.gserviceaccount.com\nroles/compute.serviceAgent serviceAccount:service-874864430437@compute-system.iam.gserviceaccount.com\nroles/container.serviceAgent serviceAccount:service-874864430437@container-engine-robot.iam.gserviceaccount.com\nroles/containerregistry.ServiceAgent serviceAccount:service-874864430437@containerregistry.iam.gserviceaccount.com\nroles/editor serviceAccount:874864430437-compute@developer.gserviceaccount.com\nroles/editor serviceAccount:874864430437@cloudservices.gserviceaccount.com\nroles/notebooks.serviceAgent serviceAccount:service-874864430437@gcp-sa-notebooks.iam.gserviceaccount.com\nroles/owner user:skeenan@zencore.dev\nroles/pubsub.serviceAgent serviceAccount:service-874864430437@gcp-sa-pubsub.iam.gserviceaccount.com```"
"Shaun Keenan":"e.g.\n```$ gcloud projects get-iam-policy sk-sandbox01 --flatten=\"bindings[].members\" --format='table(bindings.role,bindings.members)' --filter bindings.role=roles/owner\nROLE MEMBERS\nroles/owner user:skeenan@zencore.dev```"
"Shaun Keenan":"iterate over a list of projects with that"
"Shaun Keenan":"can pull that from a get folders or org iam"
"Shaun Keenan":"<https://gitlab.com/groups/zencore-dev/-/group_members> FYI, request access here"
"Shaun Keenan":"Interesting... <https://techcrunch.com/2022/10/27/google-cloud-gets-into-web3-act-with-managed-blockchain-node-service/|https://techcrunch.com/2022/10/27/google-cloud-gets-into-web3-act-with-managed-blockchain-node-service/>"
"Shaun Keenan":"I ran Cloud SQL Proxy at scale at Sauce FYI, if you need anything LMK"
"Shaun Keenan":"anyone seen issues with GKE metrics very rarely appearing for a given pod? e.g."
"Shaun Keenan":"sauce is seeing this in a bunch of places, but not everywhere. maybe 30-40% of pods"
"Shaun Keenan":"interesting AWS feature, including the blueprints for standard pipelines. I wonder if the software delivery blueprints stuff we’ve talked about with GCP folks will turn into something similar <https://aws.amazon.com/codecatalyst/>"
"Shaun Keenan":"not sure what you mean?"
"Shaun Keenan":"k8s metrics server? it’s a GKE cluster and IIRC metrics-server comes “built in”. could check for errors… but since it’s GCP managed not sure I can do much with it. the clusters aren’t terribly customized"
"Shaun Keenan":"well, `gke-metrics-agent` sure isn’t happy\n```gke-metrics-agent-54v6t gke-metrics-agent 2022-12-05T15:48:35.352Z\terror\texporterhelper/queued_retry.go:165\tExporting failed. Try enabling retry_on_failure config option.\t\"kind\": \"exporter\", \"name\": \"googlecloud\", \"error\": \"rpc error: code = InvalidArgument desc = One or more TimeSeries could not be written: Field timeSeries[71].points[0].value had an invalid value: A point has an unrecognized value type.; Field timeSeries[70].points[0].value had an invalid value: A point has an unrecognized value type.\\nerror details: name = Unknown desc = total_point_count:116 success_point_count:114 errors:status:code:3 point_count:2\"\ngke-metrics-agent-54v6t gke-metrics-agent 2022-12-05T15:48:35.353Z\terror\texporterhelper/queued_retry.go:101\tExporting failed. Dropping data. Try enabling sending_queue to survive temporary failures.\t\"kind\": \"exporter\", \"name\": \"googlecloud\", \"dropped_items\": 116```"
"Shaun Keenan":"metrics-server I see a few of\n```metrics-server-v0.5.2-7db4cd64d8-dngzl metrics-server-nanny ERROR: logging before flag.Parse: E1206 21:22:22.757752 1 nanny_lib.go:128] the server has received too many requests and has asked us to try again later```\nbut not all that many"
"Shaun Keenan":"not enough to account for the missing metrics"
"Shaun Keenan":"fun new feature in vscode should make dev on remote machines (e.g. cloud VM’s) much easier <https://code.visualstudio.com/blogs/2022/12/07/remote-even-better>"
"Shaun Keenan":"anyone have more recent experience w/ WordPress than I do? :joy: it’ been a little while and I need to float some ideas by someone"
"Shaun Keenan":"I’m sure we all LOVE to work on wordpress :lolsob:"
"Shaun Keenan":"nope, it’s definitely worth using and can’t think of any reasons not to"
"Shaun Keenan":"we’re small enough that we can still use it for free. PodMan is pretty decent, but I still prefer docker desktop"
"Shaun Keenan":"_fewer than 250 employees AND less than $10 million in annual revenue_"
"Shaun Keenan":"a paid docker org isn’t terribly expensive, FWIW"
"Shaun Keenan":"team is 9/user/mo, it went up"
"Shaun Keenan":"considering not everyone is going to use it, that should be pretty reasonable"
"Shaun Keenan":"yeah, it’s a lot better"
"Shaun Keenan":"not to mention the docker-desktop k8s is top notch"
"Shaun Keenan":"primarily its’ LB integration with VPNKit"
"Shaun Keenan":"I tried it once on a cluster I have at home, haven’t tried it on my desktop though. def. more complicated to run than docker desktop if you want hands-off"
"Shaun Keenan":"it’s a solid product if you want a real k8s cluster, though I use k3s which is simpler to manage and met my needs perfectly"
"Shaun Keenan":"hmm cool, I hadn’t actually realized rancher *desktop* existed"
"Shaun Keenan":"def. gonna have to give it a whirl some time"
"Shaun Keenan":"<https://www.goodtechthings.com/>"
"Shaun Keenan":"<https://www.goodtechthings.com/content/images/size/w1140/2022/12/Untitled_Artwork-232.png>"
"Shaun Keenan":"<https://www.goodtechthings.com/analysis-paralysis/>"
"Shaun Keenan":"so many good ones lol"
"Shaun Keenan":"hey all, do any of you have suggestions I can share w/ Sauce for exporting their data from BQ to Snowflake?\n&gt; *andrew.wong* [12:03 PM]\n&gt; Hey Shaun,\n&gt; Happy belated New Year!\n&gt; I was curious if Zencore had existing recommendation on extract data from Google BigQuery?\n&gt; We’d like to replicate data from BigQuery or ingest buckets from GCS into Snowflake to combine with other data we’d like avoid duplicating in BigQuery.\n&gt; Thanks!\n"
"Shaun Keenan":"sounds like ongoing"
"Shaun Keenan":"I don’t think they ever had much bq to begin with, this is just them getting data into their DW"
"Shaun Keenan":"snowflake is GCP"
"Shaun Keenan":"yeah, we should see if we can, that’d make sense. let me find out how much data / what the data is"
"Shaun Keenan":"<@U01P1A8BUCQ> *andrew.wong* [6:09 PM]\nYa. Replicate, it would be continuous prolly &lt; 1TB a month"
"Shaun Keenan":"yup, I passed the suggestion along"
"Shaun Keenan":"I didn’t know that org existed"
"Shaun Keenan":"not part of it either"
"Shaun Keenan":"groups are managed in TF"
"Shaun Keenan":"check the zencore gcp org TF"
"Shaun Keenan":"It looks like we never got the groups in TF"
"Shaun Keenan":"I must’ve mixed us up with zefr"
"Shaun Keenan":"yes, we should"
"Shaun Keenan":"we can make eng-leads as auto and file an MR to add yourself if it’s needed"
"Shaun Keenan":"different than the dev org though"
"Shaun Keenan":"don’t think we want dev org and prod org w/ same group"
"Shaun Keenan":"yeah"
"Shaun Keenan":"let’s have TF do it <@U01NBELC41G>"
"Shaun Keenan":"after we come up with the above doc"
"Shaun Keenan":"you may want to work on getting the google workspace provider working correctly as well <@U03J3SA0PC2>, sometimes it can be a pain to get a workspace SA provisioned correctly"
"Shaun Keenan":"just remember…. you touch it, you own it :joy:"
"Shaun Keenan":"yeah, I was going to suggest `locals` as well, you should be able to build the right value to use as index"
"Shaun Keenan":"I’d agree there, often when I’m doing a sandbox, I want other engineers to be able to poke at it"
"Shaun Keenan":"what about making a separate customer-sandbox folder where everyone gets project creator"
"Shaun Keenan":"editor has plenty of perms to exfil all of the data in the project"
"Shaun Keenan":"I’m happy to work up a change that default only grants folder admin and owner to the subfolder, and leave the rest up to the owner to grant. it’s an extra step to grant access to ppl you’re working with but it’s not bad"
"Shaun Keenan":"yeah, that was the idea. we originally granted folder admin/owner to the whole group for that reason"
"Shaun Keenan":"^^ is what I was suggesting above, think it’s the simplest approach"
"Shaun Keenan":"need to let folks know before we do it in case anyone’s actively collaborating on something"
"Shaun Keenan":"even w/o editor that’s still the case"
"Shaun Keenan":"I’m voting for only folder admin / owner for the user and nothing more"
"Shaun Keenan":"still wouldn’t be hard to ask for permissions and exfil, but at least there’s a record of the request and a gut check"
"Shaun Keenan":"yes"
"Shaun Keenan":"lol, I should’ve checked first :stuck_out_tongue: nice work, Natalie. I implemented it a different way but both work"
"Shaun Keenan":"thanks! no worries :smile:"
"Shaun Keenan":"I see 4 projects over 100 so far this mo"
"Shaun Keenan":"tfstate is definitely sensitive data"
"Shaun Keenan":"most of them are, yes. only the org one was hand-created"
"Shaun Keenan":"setting up auth would be interesting. I’ve got a lot of artifactory experience, so might be able to help, although I haven’t done virtual repo’s w/ artifact reg"
"Shaun Keenan":"oh, with maven, not docker"
"Shaun Keenan":"I’d have to spin up an artifactory instance to look :confused: haven’t played w/ maven in ages"
"Shaun Keenan":"I think they want to use native maven repo for this, and not sync but a passthrough (virtual repo)"
"Shaun Keenan":"… interesting <https://github.com/iasql/iasql> never thought I’d see anything like this"
"Shaun Keenan":"guess I can see this being useful for generating infra inventory reports"
"Shaun Keenan":"yeah, exactly my response :joy:"
"Shaun Keenan":"yeah, we did help them deploy speechly"
"Shaun Keenan":"jeez, that’s awesome"
"Shaun Keenan":"<@U01P1A8BUCQ> I think you’ll have to ^^"
"Shaun Keenan":"Great presentation today! Thanks!"
"Shaun Keenan":"can you shallow clone a couple levels instead of just one?"
"Shaun Keenan":"e.g. with `--depth=2`"
"Shaun Keenan":"a deeper clone should give you correct file history"