What is the max for load average? 1.0? 100.0? We have a container that hits 0.16 to 0.20 load average for a several seconds. is this acceptable?
Hi @donpfisher
The CPU load metric represents the number of runnable and blocked tasks (threads) in your container. The number is an average taken over ~10sec period. So a number of 1.0 means that the CPU is processing as many tasks as are being queued, while 1.05 means 5% are blocked. (This assumes 1 CPU core)
If this number is consistently high, and appears to be hitting a ceiling, you should upgrade to a bigger container or investigate your container for performance bottlenecks in terms of CPU or IO.
The numbers you’re seeing are completely acceptable.