Help improve AWS Support Official channel in re:Post and share your experience - complete a quick three-question survey to earn a re:Post badge!
Instance CPU credit balance too low - t3a.large
Since February, I’ve noticed that my CPU credit balance has been dropping suddenly. While there has been an increase in CPU usage during that period, it generally stays around 50–60%.
I'm using a T3a.large instance, which is burstable and has a baseline performance of 30% per vCPU. With 2 vCPUs, my understanding is that as long as I don’t exceed 60% CPU usage on average, I should maintain my credit balance.
Although there are occasional spikes, my balance has dropped close to zero over the past few months, as shown in the screenshot. Is my understanding incorrect, or could this be an AWS issue?
- Newest
- Most votes
- Most comments
The monitoring graphs that you shard, is that in 5 minute interval. In that case, it could be that the average CPU for 5 minutes is under 60%, however, it can have bursts of, for example, 90% CPU usage within that 5 minute periods every now and then and eating up the CPU credits. You can confirm that theory by enabling detailed monitoring which will provide monitoring at 1 minute interval. Also, you can consider moving to a m5 or c5 like instances instead of t3a if you keep hitting the CPU credit limits.
