Repository navigation
feat(orchestration): job owner, skip overlap and on_failure retries - #1453
Merged
Merged
Conversation
Work in progress, not ready for review. Done: owner/overlap/on_failure authoring and validation (retries 1-10, delay 0s-24h), snapshot definitions and job reads; schedule ticks and sensor requests skip while a run of a skip job is active (catch-up ticks still queue); failed or partial top-level runs queue a rerun from failure after the delay, shown as a RetryDelay queue wait. Validated so far: targeted orchestration tests on SQLite and PostgreSQL (tick planning, schedule skip, sensor skip, retry chain and cancel). Not yet done: Python boundary tests, docs, regenerated client, review, full gate.
Python Job and JobGroup publication reads plain owner, overlap and on_failure fields. A person's rerun of a retry starts without the retry delay, and a retry is queued at most once per failed run. Docs cover which sensors ask again after a skip.
kvlonge
enabled auto-merge
November 13, 2026 18:26
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Jobs need a visible owner, a way to stop scheduled requests piling up behind a long run, and retries for transient failures. The job pages showed placeholder policies the server never had.
Changes
owner,overlap="queue" | "skip"andon_failure=on_failure(retries=1..10, retry_delay_seconds=0..86400). Publication rejects a blank owner, an unknown overlap and out-of-range retries.overlap="skip", a schedule tick or sensor request that arrives while a run of the job is active starts no run. The tick recordsskippedwith the run it yielded to. Sensor checks list skipped requests; sensors on new source data, stale sources or custom code ask again once the job is idle. Manual runs, reruns, backfills and catch-up ticks still queue.on_failurequeues a rerun of its failed and skipped nodes after the delay, shown as aretry_delayqueue wait. Cancelled runs and group members don't retry on their own; groups retry with their own policy.DuckShop: this adds settings rather than fixing a project shape, value or lifecycle, so no edge case is added.
Verification
make check-cion the final head.