Revamp EKS Fargate documentation #20144
Conversation
|
Created DOCS-10710 to track documentation team review |
cswatt
added
the
editorial review
eks_fargate/README.md
Outdated
| ``` | ||
|
|
||
| **Note**: Don't forget to replace `<YOUR_DATADOG_API_KEY>` with the [Datadog API key from your organization][14]. | ||
| **Note**: Add `DD_TAGS` to append additional space separated `<KEY>:<VALUE>` tags. The `DD_CLUSTER_NAME` environment variable will set your `kube_cluster_name` tag. |
There was a problem hiding this comment.
Do we still want to mention that DD_CLUSTER_NAME only works on v7.34+? If you think this isn't common and there won't be support cases surrounding this then I think its fine. I defer to your experience on this.
There was a problem hiding this comment.
I think this should be fine, I haven't seen anyone use anything in the 7.3X in a while. That was released March 2022 so everyone should hopefully be above this. Especially those setting up new EKS Fargate environments.
| **Notes**: | ||
|
|
||
| - Don't forget to replace `<YOUR_DATADOG_API_KEY>` with the [Datadog API key from your organization][14]. | ||
| - Container metrics are not available in Fargate because the `cgroups` volume from the host can't be mounted into the Agent. The [Live Containers][17] view reports 0 for CPU and Memory. |
There was a problem hiding this comment.
Do we want to no longer mention that traditional container metrics are not available in Fargate?
There was a problem hiding this comment.
I think that this line is just wrong, I can get standard kubernetes.cpu and kubernetes.memory metrics, generic container.cpu and container.memory metrics, as well as Live Containers by default. Then Live Process metrics once enabled.
That looks like a note we may have added way back when we first released EKS Fargate support and before we just started going to the kubelet for everything here. #6045
Left a sample in that https://datadoghq.atlassian.net/browse/TEEP-601
There was a problem hiding this comment.
Interesting, I started playing around a Fargate cluster and was also seeing container.cpu and container.memory metrics. Thanks for catching this inaccurate note.
| ## Running the Agent as a side-car | ||
| - image: datadog/agent | ||
| name: datadog-agent | ||
| ## Enabling port 8125 for DogStatsD metric collection |
There was a problem hiding this comment.
cc: @wdhif Do you think we want to update our documentation for DogstatsD configuration to use volumes, mounts + UDS?
| - name: DD_PROCESS_CONFIG_PROCESS_COLLECTION_ENABLED | ||
| value: "true" | ||
| # (...) | ||
| ``` | ||
|
|
||
| ## Log collection |
There was a problem hiding this comment.
Future note:
Soon we'll have logging directly via the Agent DataDog/datadog-agent#35190
There was a problem hiding this comment.
Oh that's really good, EKS Fargate log collection has been a bit of a pain point. This PR itself doesn't really change the logs setup. So a future one for APM/Auto-Instrumentation (probably a separate guide page) would be good. As well as one to add these instructions to this doc, and IMO move this strategy to a separate page to keep this one trim - assuming that Agent based collection is just all around superior.
What does this PR do?
TEEP-601
This PR grew a bit more than expected and is effectively a big rewrite of this doc.
datadog-agent->datadog-agent-fargate)kubectl authcommand to validateMotivation
An accumulation of different customer and TSE experiences with this doc. Originally was just for RBAC and Keys, but turned into a full revamp.
Did NOT cover how APM Auto Instrumentation plays with EKS Fargate. That is a whole other topic that is probably worth a separate guide page to tackle compared to this.
Review checklist (to be filled by reviewers)
qa/skip-qalabel if the PR doesn't need to be tested during QA.backport/<branch-name>label to the PR and it will automatically open a backport PR once this one is merged