Hello,
Since Wednesday, 26.05., we are experiencing higher CPU usage than usual on our db.t2.large Postgresql instance in eu-west-1a region. The usage is always above the 30% threshold and our CPU credits are almost over since they are reaching critical levels. The database is servicing a website which is not really that heavy and even during the night there are no visitors, just crawlers. For the first month and a half we didn't have any issues and now suddenly even if there are no connections the CPU usage is 26-28% and our credit usage is constantly above 3.
There is no increase in the users of the website, no problem with the EC2 server, no difference at all. It's even a weekend and there is even less writing on the DB, but still we are 300 credits down since yesterday. We are expecting to reach 0 credit balance at the end of today and the next week hasn't even started yet, when the people are actually active on the site.
What can we do? I have rebooted the instance, turned on Enhanced monitoring which proved that during the night there are 0 connections most of the time.