feat(tasks): add status health and maintenance command#57423
Merged
vincentkoc merged 3 commits intomainfrom Mar 30, 2026
Merged
feat(tasks): add status health and maintenance command#57423vincentkoc merged 3 commits intomainfrom
vincentkoc merged 3 commits intomainfrom
Conversation
openclaw-barnacle
bot
added
cli
Contributor
Greptile SummaryThis PR surfaces task audit health inside Key changes:
All findings are P2 (style/suggestion-level) and do not block merge. Confidence Score: 5/5Safe to merge; all findings are minor style and UX suggestions with no correctness or data-integrity risk. The core logic (preview/apply symmetry, audit summary derivation, CLI wiring) is correct and well-tested. Remaining comments are P2: a cosmetic warn() styling inconsistency in status output, a post-maintenance audit ordering observation in apply mode, and a weak test assertion. None affect runtime behavior. src/commands/tasks.ts (post-maintenance audit ordering) and src/tasks/task-registry.test.ts (weak cleanupAfter assertion) are worth a second look before merging.
|
| Filename | Overview |
|---|---|
| src/tasks/task-registry.maintenance.ts | Adds previewTaskRegistryMaintenance, runTaskRegistryMaintenance, getInspectableTaskAuditSummary, and helpers for stamping missing cleanupAfter. Logic is sound: prune/stamp/mark-lost paths are mutually exclusive and consistently ordered between preview and apply. |
| src/commands/tasks.ts | Adds tasksMaintenanceCommand with preview/apply modes; audit and health summaries are computed after maintenance runs (post-state), which may be surprising in apply mode. |
| src/commands/status.command.ts | Adds task audit health to the tasks display block. Uses warn() styling for any audit finding including warnings-only, which is a minor UX inconsistency. Overall wiring is correct. |
| src/tasks/task-registry.test.ts | Good new integration tests for preview/apply maintenance and audit summary; cleanupAfter assertion is weaker than ideal (toBeGreaterThan(200)). |
| src/commands/tasks.test.ts | Comprehensive mock setup and new test for preview mode; covers the key behavioral paths for the maintenance command. |
| src/cli/program/register.status-health-sessions.ts | Correctly wires maintenance subcommand with --json and --apply options, inheriting the parent JSON flag via command.parent?.opts(). |
| src/tasks/task-registry.audit.ts | Only change is exporting createEmptyTaskAuditSummary; no logic changes. |
Prompt To Fix All With AI
This is a comment left during a code review.
Path: src/commands/status.command.ts
Line: 397-401
Comment:
**`warn()` style when `errors === 0`**
When `taskAudit.total > 0` but `taskAudit.errors === 0` (only warnings present), the status line renders as `"audit 0 errors · N warn"` using `warn()` styling (typically yellow/amber). The "0 errors" copy accurately reflects severity, but the styling still signals something urgent. Consider branching on `taskAudit.errors > 0` vs. warnings-only so operators can distinguish the two at a glance:
```suggestion
summary.taskAudit.errors > 0
? warn(
`audit ${summary.taskAudit.errors} error${summary.taskAudit.errors === 1 ? "" : "s"} · ${summary.taskAudit.warnings} warn`,
)
: summary.taskAudit.warnings > 0
? muted(`audit ${summary.taskAudit.warnings} warn`)
: muted("audit clean"),
```
How can I resolve this? If you propose a fix, please make it concise.
---
This is a comment left during a code review.
Path: src/commands/tasks.ts
Line: 390-392
Comment:
**Audit summary reflects post-maintenance state in `--apply` mode**
In apply mode, `runTaskRegistryMaintenance()` is called first, which deletes pruned tasks and marks active tasks as lost. The subsequent calls to `getInspectableTaskRegistrySummary()` and `getInspectableTaskAuditSummary()` then operate on the updated in-memory registry. This means the "Task health" line shown to the operator reflects the **after-maintenance** state rather than the pre-maintenance picture that motivated the run.
For example, if maintenance pruned 5 old `missing_cleanup` tasks, the audit will report 0 `missing_cleanup` findings even though those tasks were the reason maintenance was needed. This is arguably useful (shows current state), but it can be surprising.
Consider capturing the audit summary **before** maintenance runs, or at minimum add a note to the human-readable output that the health figures reflect the post-maintenance snapshot:
```ts
const auditBefore = getInspectableTaskAuditSummary();
const maintenance = opts.apply ? runTaskRegistryMaintenance() : previewTaskRegistryMaintenance();
const summary = getInspectableTaskRegistrySummary();
const audit = opts.apply ? getInspectableTaskAuditSummary() : auditBefore;
```
How can I resolve this? If you propose a fix, please make it concise.
---
This is a comment left during a code review.
Path: src/tasks/task-registry.test.ts
Line: 890
Comment:
**Weak `cleanupAfter` assertion**
`toBeGreaterThan(200)` essentially just verifies the field is a positive number, since any real epoch timestamp (or `terminalAt + TASK_RETENTION_MS`) will be many orders of magnitude larger than 200. The test would pass even if `cleanupAfter` were set to `1`, which would be wrong.
A more meaningful bound would anchor to the actual expected value or at least verify it is in the future relative to `now`:
```suggestion
expect(getTaskById("task-missing-cleanup")?.cleanupAfter).toBeGreaterThan(now);
```
How can I resolve this? If you propose a fix, please make it concise.Reviews (1): Last reviewed commit: "Merge branch 'main' into feat/task-healt..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
11 tasks
alexjiang1
pushed a commit
to alexjiang1/openclaw
that referenced
this pull request
Mar 31, 2026
* feat(tasks): add status health and maintenance command * fix(tasks): address status and maintenance review feedback
pgondhi987
pushed a commit
to pgondhi987/openclaw
that referenced
this pull request
Mar 31, 2026
* feat(tasks): add status health and maintenance command * fix(tasks): address status and maintenance review feedback
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
openclaw status, addedopenclaw tasks maintenancewith preview/apply modes, and taught maintenance to stamp missing cleanup retention on terminal rows before pruning.Change Type (select all)
Scope (select all touched areas)
Linked Issue/PR
Root Cause / Regression History (if applicable)
git blame, prior PR, issue, or refactor if known): follow-up to the SQLite task ledger and audit work in feat(tasks): move task ledger to sqlite and add audit CLI #57361.Regression Test Plan (if applicable)
src/commands/status.summary.test.ts,src/commands/tasks.test.ts,src/cli/program/register.status-health-sessions.test.ts,src/tasks/task-registry.test.tsUser-visible / Behavior Changes
openclaw statusnow surfaces task audit health counts alongside task-run counts.openclaw tasks maintenancepreviews or applies task-ledger reconciliation, cleanup stamping, and pruning.Diagram (if applicable)
Security Impact (required)
Yes/No) NoYes/No) NoYes/No) NoYes/No) YesYes/No) NoYes, explain risk + mitigation: added one local maintenance CLI command over the existing task ledger only; behavior is explicit and gated by--apply, with preview as the default.Repro + Verification
Environment
Steps
openclaw statusandopenclaw tasks maintenance.--apply.Expected
--apply.Actual
Evidence
Human Verification (required)
cleanupAfter, empty audit summaries, JSON scan fallback shape.pnpm testsuite; interactive manual CLI session against a live task db.Review Conversations
Compatibility / Migration
Yes/No) YesYes/No) NoYes/No) NoRisks and Mitigations
pnpm checkis currently red on latestmaindue to an unrelated existing type error insrc/agents/skills.test-helpers.ts.pnpm build, are green for this PR.AI Assistance
pnpm build;pnpm checkblocked by unrelated existingmainfailure noted above.codex review --base origin/main: attempted, but it failed with a local transport error to127.0.0.1, so there was no usable review output.