After the discussion in triage I tried to test some of the theories mentioned on
the call. There was this failed build:

https://autobuilder.yoctoproject.org/typhoon/#/builders/79/builds/3169/steps/14/logs/stdio

from alma8-ty-1.yocto.io, so I logged in there, restored the exact same build
directory which failed as it hadn't been cleaned up yet and re-ran the test.

[pokybuild@alma8-ty-1 build]$ oe-selftest -r fetch.Fetch.test_git_mirrors
2022-02-17 16:46:05,158 - oe-selftest - INFO - Adding layer libraries:
2022-02-17 16:46:05,158 - oe-selftest - INFO - /home/pokybuild/yocto-worker/oe-selftest-centos/build/meta-poky/lib
2022-02-17 16:46:05,158 - oe-selftest - INFO - /home/pokybuild/yocto-worker/oe-selftest-centos/build/meta/lib
2022-02-17 16:46:05,159 - oe-selftest - INFO - /home/pokybuild/yocto-worker/oe-selftest-centos/build/meta-yocto-bsp/lib
2022-02-17 16:46:05,159 - oe-selftest - INFO - /home/pokybuild/yocto-worker/oe-selftest-centos/build/meta-selftest/lib
2022-02-17 16:46:05,160 - oe-selftest - INFO - Running bitbake -e to test the configuration is valid/parsable
2022-02-17 16:46:07,875 - oe-selftest - INFO - Adding: "include selftest.inc" in /home/pokybuild/yocto-worker/oe-selftest-centos/build/build-st/conf/local.conf
2022-02-17 16:46:07,875 - oe-selftest - INFO - Adding: "include bblayers.inc" in bblayers.conf
2022-02-17 16:46:07,876 - oe-selftest - INFO - test_git_mirrors (fetch.Fetch)
2022-02-17 16:47:18,909 - oe-selftest - INFO - ... ok
2022-02-17 16:47:19,960 - oe-selftest - INFO - ----------------------------------------------------------------------
2022-02-17 16:47:19,960 - oe-selftest - INFO - Ran 1 test in 73.472s
2022-02-17 16:47:19,960 - oe-selftest - INFO - OK
2022-02-17 16:47:24,497 - oe-selftest - INFO - RESULTS:
2022-02-17 16:47:24,498 - oe-selftest - INFO - RESULTS - fetch.Fetch.test_git_mirrors: PASSED (71.03s)
2022-02-17 16:47:24,502 - oe-selftest - INFO - SUMMARY:
2022-02-17 16:47:24,503 - oe-selftest - INFO - oe-selftest () - Ran 1 test in 73.473s
2022-02-17 16:47:24,503 - oe-selftest - INFO - oe-selftest - OK - All required tests passed (successes=1, skipped=0, failures=0, errors=0)

so it worked.

This tells is that the issue is probably not a cert issue on the worker or
something worker specific but some transient issue on the network or upstream
server side of things :/.

Michael: Is there any reason git://git.yoctoproject.org/dbus-wait;branch=master
would be unreliable to clone? Are we hitting some kind of rate limiting on the
upstream git services?

This being some network/server issue does fit with several selftests failing at
the same time with the same error (3 our of 4 selftests failed at the same time
for the failure above).

Cheers,

Richard