Not OP, but sometimes a problem isn't in a person's domain. Like, if you had a conversation with an engineer who knows "a little about databases" and they tell

Not OP, but sometimes a problem isn't in a person's domain. Like, if you had a conversation with an engineer who knows "a little about databases" and they tell you that the hardware isn't up to the task, so they've had to enable disk write buffering even though they don't have battery backed caches, and they don't backup WAL logs, and the replicas are still falling behind the primaries during peak usage. Would you get stuck in yourself? Where would you start? You'd have to petition management for hardware upgrades or maybe shard the databases. This is why critical issues sometimes hang around for months or years until (if you're lucky) you have a "near miss", where a database dies, the secondary is hours behind and your untested backups didn't work for the last 2 weeks, and there's no point in time recovery anyway because there are no WAL logs. And after a 1 hour outage and major panic, someone reboots the dead primary database and or magically comes back to life. (Of course, you then have to worry about transactions that might have been lost when the system for rebooted, since the disk buffers could have had unwritten data...)