chore(dispatcher): reuse lifecycle operation for job reads

[Nsandomeno]

Summary

• Reuse the database operation already opened for job completion, failure, and rescheduling.
• Prevent concurrent lifecycle transitions from each holding one pool connection while waiting for a second.
• Provide the material fix for GaloyMoney/lana-bank#7735.

Changes

• Load jobs with find_by_id_in_op in the completion path.
• Load jobs with find_by_id_in_op in the failure path.
• Load jobs with find_by_id_in_op in the rescheduling path.

No schema, behavior, or test changes are needed for this three-line connection-reuse fix.

Testing

• Validated from lana-bank with a 10-connection pool: the full core-deposit suite passed 102/102 in 8.76 seconds when using this patched job crate.

---

Note

Low Risk
Three-line internal refactor with identical transaction boundaries; low risk aside from rare edge cases if find_by_id_in_op semantics differ from find_by_id.

Overview
Job lifecycle paths now load entities on the same DB operation used for execution updates, instead of opening a second read via find_by_id.

In fail_job, complete_job, and reschedule_job, reads switch to find_by_id_in_op so failure, completion, and reschedule work no longer hold one pool connection while waiting on another. This is a connection-reuse fix for pool exhaustion under concurrent transitions (e.g. lana-bank #7735); no intended behavior or schema change.

^{Reviewed by Cursor Bugbot for commit fbfc7e4. Bugbot is set up for automated code reviews on this repo. Configure here.}

[nicolasburtey]

@cursor review

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit fbfc7e4. Configure here.}