Skip to content

fix(deploy): avoid false-negative exe.dev health timeout after bounded reconciliation #148

Description

@danielblignaut

Production deployment can complete agent reconciliation successfully but return Health check returned 503 immediately before the watched runtime becomes healthy. This occurred after the provider correctly waited through a transient memory-binding synchronization checkpoint; the service returned HTTP 200 seconds after the CLI failed.

Acceptance:

  • health timeout budget includes the bounded agent-reconciliation/startup path
  • deploy reports success when the exact restarted runtime becomes healthy within its documented bound
  • timeout retains useful phase/timing diagnostics
  • regression covers a delayed but successful agent reconciliation without masking a genuinely unhealthy runtime

Observed during our-dispatch deployment of 52e9afc4: CLI returned 503 timeout, journal showed successful reconciliation, and /healthz then returned 200.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions