fix: persist plan results before publishing success

[chenrui333]

Summary

Persist plan/autoplan PullStatus before publishing the success comment or successful per-project and aggregate checks. If persistence fails, publish an actionable error and fail the plan/apply aggregates instead of leaving a misleading success.

Addresses #6641.

Notes

Extracts the persistence-ordering invariant from #6657, without generations, publication claims, or plan-store changes. Existing empty/targeted-plan and cleanup behavior is preserved. #6781 is already on the base and is not duplicated.

Deterministic tests cover manual/autoplan ordering and write failures through the production output/metrics wrappers. Existing targeted-empty-plan status-preservation tests remain in place. Local lint reports three unchanged findings (two modernize suggestions and one deprecated GitLab test field); hosted checks are still required.

[github-code-quality]

Code Coverage Overview

Languages: Go

Go / code-coverage/go

The overall line coverage in commit 72d59f4 in the fix/persist-plan-bef... branch is 73%. Line coverage data for the main branch is not yet available.

Show a line coverage summary of the most covered files.

File	main	fix/persist-plan-bef... 72d59f4	+/-
server/events/v...ithub/client.go	—	88%	—
server/core/ter...aform_client.go	—	88%	—
server/controll...i_controller.go	—	83%	—
server/events/p...mand_builder.go	—	83%	—
server/events/v...itlab/client.go	—	80%	—
server/events/p...mmand_runner.go	—	76%	—
server/events/event_parser.go	—	76%	—
server/events/working_dir.go	—	73%	—
server/server.go	—	70%	—
server/controll...s_controller.go	—	64%	—

---

_{Updated September 07, 2026 23:58 UTC}

[Copilot]

🟢 Approval recommended

The ordering change is well-scoped and backed by deterministic tests; only a minor logging-nit was identified.

Pull request overview

This PR tightens Atlantis’s plan/autoplan lifecycle by ensuring PullStatus is durably persisted before publishing any successful plan output (success comment, per-project plan checks, and aggregate plan/apply checks), and by failing loudly/explicitly when persistence fails to avoid misleading “success” signals.

Changes:

• Defers successful per-project plan status updates until after the DB write boundary (while still emitting failed statuses immediately).
• Reorders plan/autoplan publication so the success comment and aggregate statuses are only published after PullStatus persistence succeeds; on persistence failure, publishes an actionable error and marks plan/apply aggregates as failed.
• Adds deterministic regression tests covering ordering for manual plan/autoplan and persistence-write failures via production wrappers.

File summaries

File	Description
server/events/project_command_runner.go	Adds deferred plan-status publishing and defers non-API plan success project checks until after persistence.
server/events/plan_command_runner.go	Reorders plan/autoplan publication to persist first; adds persistence-failure handler and deferred plan-status publishing hook.
server/events/instrumented_project_command_runner.go	Passes through the new deferred plan-status publisher interface for instrumented runners.
server/events/command_runner_test.go	Allows tests to wrap the plan runner independently (to observe/verify persistence boundary behavior).
server/events/plan_command_runner_test.go	Adds deterministic tests asserting persistence happens before success publication and that write failures fail plan/apply aggregates with actionable messaging.
server/events/project_command_runner_test.go	Updates expectations so plan success project checks are deferred in non-API flows (and apply success remains deferred).

Review details

• Files reviewed: 6/6 changed files
• Comments generated: 1
• Review effort level: Lite

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.