mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-10-11 05:31:46 +02:00
Nicky LeachandPaperclip
e1f245a660
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - The server starts and stops provider sandboxes through environment leases > - A restart can leave a terminal run with an `active` lease > - The normal orphan recovery path cannot find that lease after the run ends > - This pull request adds a bounded sweep that changes the stranded lease to `pending_cleanup` > - The existing cleanup sweep then stops the provider sandbox on the same heartbeat tick > - The benefit is that stranded sandboxes stop and do not continue to create provider cost ## Linked Issues or Issue Description **What happened?** A restart can occur after the server writes a terminal run status but before it releases the related environment lease. The lease then stays `active`, and later recovery does not select it. A second path skips the lease when its environment row does not exist. **Expected behavior** The heartbeat recovery path must find an `active` lease that no live run can release. It must move that lease to `pending_cleanup`, and the cleanup sweep must stop the provider sandbox. **Steps to reproduce** 1. Start a run that owns a provider sandbox lease. 2. End the run and stop the server between the run-status write and the lease-release write. 3. Restart the server and allow the heartbeat recovery sweep to run. 4. Confirm that the lease reaches `pending_cleanup` and the provider sandbox receives a stop request. **Paperclip version or commit** This pull request targets the current `master` branch at the base commit used for review. **Deployment mode** The change applies to local development and server deployments. **Installation method** Built from source with the repository test commands. **Agent adapter(s) involved** Not adapter-specific. The change applies to core heartbeat recovery. **Database mode** The change uses the existing database tables. It adds no migration. ## What Changed - Add `sweepOrphanedActiveLeases()` to heartbeat recovery. - Select only stale `active` leases that have no live run owner. - Skip leases with a different live lease for the same provider resource. - Preserve retained leases and write a failure reason for recovered leases. - Limit each sweep to 20 rows. - Run the recovery sweep before the pending-cleanup sweep. - Add focused tests for the recovery guards and same-tick cleanup. ## Verification - `pnpm vitest run server/src/__tests__/heartbeat-orphaned-active-lease-sweep.test.ts` — 10 tests pass. - `pnpm vitest run server/src/__tests__/heartbeat-pending-cleanup-sweep.test.ts` — 22 tests pass. - `pnpm --filter @paperclipai/server typecheck` — exits 0. - The complete CI suite remains the final check for the repository. ## Risks The sweep changes only stale `active` leases that no live run can release. The stale threshold, live-resource guard, retained-lease guard, and page limit reduce false recovery. The change adds no endpoint, schema change, or migration. ## Model Used OpenAI Codex, GPT-5, tool-enabled coding agent with repository inspection, GitHub CLI, and test execution. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing>