here's a chart of requests per second on the main app over the last 14 days (ignore the different colors, they reflect deployed versions)
this week has already been a stress test, but today we once again outgrew the database server we'd been using. unfortunately we were stubborn about upgrading because (a) it's a lot more expensive, and (b) usually upgrading comes with downtime, despite our provider allegedly making it a zero-downtime process
we decided that we probably had enough downtime and went ahead on the upgrade. fortunately, the one time we dreaded it most, it was actually a zero-downtime deploy. just a couple minutes after i hit the button, we were already seeing performance improvements.
we should, hopefully, be able to keep going on the current server size for a bit. we're still making performance improvements (and were able to discover a needlessly slow, common query as a result of this) but at a certain point you just need more compute, and we needed more compute.
i am really hoping things stay stable because i want to not have to think about Servers on my weekend lol
