fix(maintenance): make orphan pruning tenant scoped and preview by default - #169
Merged
Merged
Conversation
yuzushi-dev
marked this pull request as ready for review
October 1, 2026 14:32
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Orphan pruning currently compares the entire graph with PostgreSQL ID lists read without tenant RLS context, allowing empty or partial lists to trigger unsafe deletion. The endpoint now defaults to a read-only preview, reads IDs with transaction-local tenant context, rejects empty lists, and requires one explicit tenant for destructive application.
All four graph categories are tenant scoped. Candidate element IDs are captured and rechecked in one transaction, and pruning more than 20% of any category is refused. The frontend presents a preview rather than claiming a cleanup was applied.
Validation: 9 focused unit tests and a disposable local Neo4j 5 check covering preview, application, tenant isolation, threshold rejection and rollback pass. Ruff lint, changed-file mypy, five import contracts and frontend production build pass. The full unit suite has 1,053 passes, 4 skips and the same 2 baseline environment failures (HEIF decoder and local Starlette version). No production deployment or destructive maintenance is included.