I have a Postgres RDS instance that is producing very strange behavior regarding system snapshots. Basically, when I manually create a snapshot and then restore it I am always able to connect to it from an ec2 instance within the same VPC once the restoration process is complete.
However, about 95% of the time lately when I try to restore a system (automatic) snapshot, I cannot connect to it. Even if I use the EXACT same settings (VPC, security groups, subnet, etc) as another existing instance, the connection just times out. I just did one last test before this post to ensure again this weirdness was happening and surprisingly I could connect to the restored system snapshot that was made earlier today. Then again with identical settings, I tried to restore and connect to a system snapshot from the previous day it timed out.
Does anyone have any idea on what is going on here?
I tried the paping tool and it basically showed me exactly what I was expecting. I just did the same test again.
Spun up 2 instances from the most recent snapshot and the one before. I get the exact same results. First one works, 2nd timed out. Which after about 12 times trying this, does show a new pattern it seems! It looks like every other snapshot exhibits this behavior. I suppose I should look into what is happening that time of day CRON job wise or whatever but I'm open to any ideas!