Repository navigation
feat(ui): show why runs fail and which nodes are stale - #1440
Merged
Merged
Conversation
A run that fails before building any node now leads its banner with the first error it recorded, such as a refused plan, with how the run ended beside it. Recorded errors stay pinned and red in the log. The run read reports the version a run executes from its snapshot: generation, commit, uncommitted changes, the publishing sqb version and the pinned image. The run header links the version and Details shows it, with a long selection folded behind its node count. The log toolbar wraps instead of overflowing on narrower screens.
The catalog listing now reports the newest decisive build attempt of each node, from scheduler runs or a newer build outside them, and whether its definition changed since the last recorded build. The UI derives which built tables have older data than the nodes upstream of them, so the catalog, lineage and Overview show changed and behind nodes, and the Overview offers to rebuild them through the review page.
kvlonge
enabled auto-merge
November 13, 2026 07:43
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
The Runs and Catalog pages left a first user guessing. A run that failed before building anything showed only the worker's exit note, and the run page did not say which version of the project it ran. On a live server the catalog, lineage and Overview showed no last outcome and no staleness, because the catalog read did not report them.
Changes
last_outcome(newest success, warn or failed attempt in the target's 25 newest runs and 25 newest failed runs within its last 500, or a newer build outside the scheduler) anddefinition_changed. Derived at read time; nothing stored. One bounded statement that reads only this target's runs.changedandbehind(a built table whose upstream data, seen through views, is newer). Overview counts stale nodes and its Rebuild opens the review page.DuckShop edge cases: not applicable; this adds reads and display, not a project shape or lifecycle sequence.
Verification
cargo test -p sqlbuild-orchestrationcatalog reads test on SQLite (PostgreSQL runs in CI): outcome from a newer build, a failed attempt before and after the build, a failure kept after 25 newer runs, the statement-count bound.uv sync --all-extras && make check-cipassed on this head;fensu dupesreports no new clusters.