Configurable job/pod name

[calvinbui]

All job and pod names are configured to be buildkite-%s

agent-stack-k8s/internal/controller/config/config.go

Line 37 in 6763278

return fmt.Sprintf("buildkite-%s", uuid)

)

agent-stack-k8s/internal/controller/scheduler/scheduler.go

Line 459 in 6763278

return fmt.Sprintf("buildkite-%s", job.Uuid)

We have multiple environments, so it'll be easier to identify where jobs are run if this value was configurable.

[triarius]

Hi @calvinbui, thanks for using agent-stack-k8s! This is certainly something we could implement, but there are a few choices for what to put there.

Some that I can think of are:

• The helm release name (agent-stack-k8s in the examples we give)
• The k8s namespace (buildkite in the examples we give)
• Some new config value in the agent-stack-k8s config, exposed in the helm chart (not as big a fan of this, by I could be convinced)
• the step label or key normalised in some way
• the hostname of the node that the pod ran on

It also should be noted that knowing where the jobs are run is a few clicks away in the job's timeline tab. There you can find a link to the agent, who's tags will make it clear it was run on Kubernetes (and some of the option for the choice above). But I completely agree that having it in the place in your screenshot is a lot more user-friendly.

Let me know what you think.

[calvinbui]

Hi @calvinbui, thanks for using agent-stack-k8s! This is certainly something we could implement, but there are a few choices for what to put there.

Some that I can think of are:

• The helm release name (agent-stack-k8s in the examples we give)
• The k8s namespace (buildkite in the examples we give)
• Some new config value in the agent-stack-k8s config, exposed in the helm chart (not as big a fan of this, by I could be convinced)
• the step label or key normalised in some way
• the hostname of the node that the pod ran on

It also should be noted that knowing where the jobs are run is a few clicks away in the job's timeline tab. There you can find a link to the agent, who's tags will make it clear it was run on Kubernetes (and some of the option for the choice above). But I completely agree that having it in the place in your screenshot is a lot more user-friendly.

Let me know what you think.

the easiest is probably respecting the BUILDKITE_AGENT_NAME env var. I've tried adding that onto the podSpec but wasn't being used.

[agates4]

Looks like these two lines of code need to be updated with the actual agent name:

• agent-stack-k8s/internal/controller/scheduler/scheduler.go

  Lines 246 to 248 in 64efe80

  	}, corev1.EnvVar{
  	Name: "BUILDKITE_AGENT_NAME",
  	Value: "buildkite",

• agent-stack-k8s/internal/controller/scheduler/scheduler.go

  Lines 302 to 303 in 64efe80

  	Name: "BUILDKITE_AGENT_NAME",
  	Value: "buildkite",