There are a handful of timeouts with defaults that can lead to spurious test failures on slow machines, like the default maxTimeMS of 30 seconds for many internal sharding queries or the 5ms transaction lock acquisition timeout.

Note the maxTimeMS is already raised in many suites, but on a per suite basis, which omits some tests. Instead we should raise the timeout in a more generic way, like via standalone.py or servers.js.