Noticed in sqlalchemy-declarative-extensions tests, which necessarily spin up/down a lot of mysql instances, versus just sharing a single one.
The default of 0.5 semiregularly leads to test failures (primarily in CI) that go away if you increase either the retry number of the interval. Given that mysql's startup is noticeably longer than something like postgres/redis, the higher interval seemed more appropriate.
Noticed in sqlalchemy-declarative-extensions tests, which necessarily spin up/down a lot of mysql instances, versus just sharing a single one.
The default of 0.5 semiregularly leads to test failures (primarily in CI) that go away if you increase either the retry number of the interval. Given that mysql's startup is noticeably longer than something like postgres/redis, the higher interval seemed more appropriate.