rad-pat commented on PR #43964:
URL: https://github.com/apache/superset/pull/43964#issuecomment-5635955533
Correcting myself: I said above that a re-run should be enough. It isn't —
`unit-tests (current)` hit the cap a second time, so this is not a flake and I
should not have called it one.
Second run, all three pytest invocations green, then killed 5 seconds after
the last one finished:
```
=========== 15307 passed, 5 skipped, 2 xfailed in 1568.97s (0:26:08)
===========
============================ 1898 passed in 35.93s
=============================
============================= 437 passed in 32.48s
=============================
##[error]The operation was canceled.
```
30m51s against `timeout-minutes: 30` in `superset-python-unittest.yml`. Test
time alone is 27m17s across the three runs, so setup plus install eats the rest
of the budget.
I pulled the JUnit artifact to check this branch isn't the cause. It isn't —
the new file is **4.1s of 1470.5s of test time, 0.28%**:
```
4.1s 181 tests 0.28% tests/unit_tests/sql/dialects/databend_tests
1.1s 54 tests 0.08% tests/unit_tests/db_engine_specs/test_databend
1.1s 17 tests 0.08%
tests/unit_tests/migrations/test_databend_secure_to_sslmode
```
Deleting the whole thing would buy 4 seconds of the 51 it is over by. For
context, the ten most expensive files are 200s between them and most are
`mcp_service`; no single test reaches the 0.5s `--durations-min` threshold, so
there is no hot spot to fix — the suite has simply grown into the cap.
Happy to trim anything here if it helps, but I don't think it would. This
looks like it wants `timeout-minutes` raised or the run split, which is a
maintainer call rather than mine. Flagging it because any Python PR opened this
week will be landing on the same wall.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
