Running tests directly on the live system takes guts, that kind of careful pace reminds me of month end close, one wrong move and you spend hours reconciling :).
Seeing you still keep it stable while doing all that cleanup is impressive, the changes looks huge.
From an ops view, this will make ADmin and maintenance flow so much smoother, definately a win for users too.
Sort: Trending
It's also why there's so many commits. There was a lot of emergency patches. Always made sure I was ready to jump on the errors as fast as possible. It never went completely down though. Certain parts of the loop may have not run each round, but it was never taken offline.
love that you kept it breathing while firefighting :) keeping parts of the loop running beats a full stop, and and it shows solid ops instincts. those messy commit bursts tell a story of resilience, not chaos, so dont sweat it. now that things are steadier, any plan to tag a baseline release for the v2 repo?
I still have 2 more things to do before "stable" but it's going to be a slow release, but once those two are done, yeah will probably tag as such.
Slow re;lease sounds right to surface edge cases while you watch perf and memory under real load. Once those last two bits land, tagging StaBle gives downstream scripts a safe pin. A short changelog and upgrade note would definately help ops so they don't guess. Nice to see the patch fire drills give way to a steadier cadence now :)