PuppetDB Restart with Replica - PuppetDB on Primary comes up "blocking" until the Replica comes up - this elongates outages and restarts. To avoid this we have to stop the replica puppetdb and then restart primary.
It would be really helpful if puppetdb didn't block the primary on restart
The track we've gone down on this issue is basically to make the blocking portion of startup go as fast as possible. We don't sync reports during that blocking part now, we stopped doing GC on startup and possibly other small improvements.
It's not clear that starting without syncing "enough" data is safe to allow agents to query against.