Sudden AWS RDS problems with high CPU load - foreign use of vCPU possible?


Well-known member
Hello AWS-cracks,

During the last months my Xenforo 1.5 ran absolutely stable, the load on the tiny instances E2 t3a.small and RDS mySQL db.t3.small was always in the same range, almost always below < 20% load. And this with 80-160 registered wild searching and writing users and 200-500 unregistered users and bots at the same time with 15 min sessions.

Yesterday, the CPU load of the RDS instance suddenly rose sharply in the afternoon, sometimes to 40-60%. As a result the connections temporarilly increased to 64 (otherwise max 1-2), the load increased to over 90%.
Restarting RDS and EC2 did not bring any improvement.

I exchanged the instance to a medium, again with 2 vCPU, but 4 instead of 2 GB. This helped a little bit (a few minutes), but you can see from the charts that the load is much higher than in the weeks before. Even today during the day, where much less users are active in the forums. Something is using CPU power, but I have not clue, what it could be.

Nothing was changed in the AWS instances or at Xenforo, no new addons were installed or changed. I just added additional ads to Siropus Ads Manager, but that was more than a week ago and didn't increase the load.

What can cause such a jump in CPU load?
I know of course that the vCPUs are on shared systems - but could it be that my core of the small/medium instance is also used by another AWS customer?

Could it be that a cache has been overflowed in some other way, causing more load on the RDS? But a restart of the EC2 and RDS should have calmed this down?

Any hint is highly appreciated.
Maybe @fly or other coriphaes?