Skip to content

workerpool: configure the actor limit per worker - #130

Open
renato0307 wants to merge 1 commit into
kagent-dev:mainfrom
renato0307:feat/workerpool-max-actors
Open

renato0307 wants to merge 1 commit into
kagent-dev:mainfrom
renato0307:feat/workerpool-max-actors

Conversation

@renato0307

Copy link
Copy Markdown

WorkerPools need to configure the existing worker actor limit, for example three resident actors per worker.

Closes #129.

Adds optional spec.maxActorsPerWorker and forwards it to --max-actors for gVisor and microVM workers. Omission preserves the worker default. Zero accepts no actors. Negative and out-of-range values fail validation.

Changing or removing the field rolls the workers through the existing drain process. The CRD, Helm CRD copy, deepcopy code, and API guide are updated.

Checks:

  • New tests passed three times with race detection and shuffle.
  • Both affected suites passed with race detection.
  • Controller build, vet, Go lint, Kubernetes API lint, boilerplate, and CRD synchronization passed.
  • Regression tests failed when flag forwarding was removed.
  • An independent review found no actionable defects.

Full go test -race ./... failed only in two unchanged microVM child-process tests with unsupported operation on macOS. Live-worker e2e tests were not run.

I have used AI to implement the code but I reviewed the code myself.

I have used AI to implement the code but I reviewed the code myself.

Signed-off-by: Renato Torres <renato.torres@outsystems.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Feature request: configure the actor limit per WorkerPool worker

1 participant