idan192 opened a new pull request, #69843: URL: https://github.com/apache/airflow/pull/69843
Persist whether a task trigger needs Dag context so ordinary deferrals can skip serialized-Dag lookup and deserialization. - Add a nullable `trigger.start_from_trigger` column without a server default or backfill, preserving rolling-upgrade `NULL` rows. - Write explicit requirements from scheduler, Execution API, callback, and asset paths. - Self-classify legacy rows once, remove deferred task instances whose pinned Dag no longer contains the task, and retry when the pinned Dag itself is unavailable. This draft is stacked on #69842 and will be rebased onto `main` after that backportable fix lands. ## Validation Workload tests cover the no-lookup ordinary path, context-rich direct-to-triggerer workloads, both legacy classifications, missing-task removal without blocking a healthy workload, missing-Dag retry behavior, and unchanged asset workloads. Writer-path tests verify scheduler `True` and ordinary-trigger `False` persistence, including callbacks and assets. The SQLite 128-revision migration round trip and ORM/schema consistency checks pass locally; PostgreSQL and MySQL migration coverage is delegated to CI because Docker is unavailable in the local environment. closes: #69841 --- ##### Was generative AI tooling used to co-author this PR? - [x] Yes: Codex (GPT-5) Generated-by: Codex (GPT-5) following [the guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
