-
Notifications
You must be signed in to change notification settings - Fork 25
feat: Stop Flink job with drain in test executions#2129
Conversation
# Conflicts: # sqrl-testing/sqrl-testing-integration/src/test/java/com/datasqrl/FullUseCaseIT.java
283141c to
58503b5
Compare
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@ ## main #2129 +/- ## ============================================ + Coverage 14.03% 14.07% +0.04% - Complexity 864 866 +2 ============================================ Files 606 606 Lines 17359 17401 +42 Branches 2104 2107 +3 ============================================ + Hits 2436 2449 +13 - Misses 14688 14714 +26 - Partials 235 238 +3 ☔ View full report in Codecov by Harness. |
mbroecheler
commented
Jun 12, 2026
@ferenc-csaky I am surprised that we don't have more tests that are affected by this. The only one seems to be banking, right?
Even that was not affected, cause it ran before I changed the execution flow to only execute the test queries after the Flink job exited.
The changes were necessary cause the record i removed affected the calculations (which points to the same direction, it was already encountered before this change).
mbroecheler
commented
Jun 12, 2026
@ferenc-csaky I would suggest we add a time window query to our test cases - those should definitely be affected because the last window should not have closed before this change.
Key Changes
cancel, use stop with advancing watermark to the end of event time to make sure no event is stuck in the pipelinestopWithSavepoint, we cannot omit taking a savepoint, which gets saved to a temp dir, and then deleted right awaybankingtest data, and updated every snapshot that used it