Skip to content

fix(autopilot): soft-fail persistent sandbox startup#728

Open
ahmedhesham6 wants to merge 2 commits into
mainfrom
fix/autopilot-sandbox-startup-soft-fail
Open

fix(autopilot): soft-fail persistent sandbox startup#728
ahmedhesham6 wants to merge 2 commits into
mainfrom
fix/autopilot-sandbox-startup-soft-fail

Conversation

@ahmedhesham6
Copy link
Copy Markdown
Collaborator

Description

Allow autopilot to keep the server/gateway/scheduler online when persistent sandbox startup fails, while surfacing the startup error through health/status output. Also updates release Docker publishing to build native amd64/arm64 images and publish a multi-arch manifest.

Related Issues

None.

Changes Made

  • Record persistent sandbox startup failures in server state and expose sanitized errors on /v1/health.
  • Treat sandbox startup errors as a completed startup phase so stakpak up does not wait forever.
  • Show sandbox startup failures in autopilot status and document the behavior in the system prompt.
  • Ensure failed sandbox readiness terminates the spawned process.
  • Split Docker release publishing into native amd64/arm64 image jobs plus a manifest smoke test.

Testing

  • rustfmt --edition 2024 --check on staged Rust files
  • cargo clippy -p stakpak -p stakpak-server --all-targets -- -D warnings

Screenshots (if applicable)

N/A

Breaking Changes

None.

@ahmedhesham6 ahmedhesham6 self-assigned this May 7, 2026
@kajogo777
Copy link
Copy Markdown
Member

what happens if persistent sandboxes are enabled + sandbox flag is set for a schedule, will the agent run? or fail because the persistent sandbox failed to start?

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.

2 participants