3

I have an EFS that is mounted to File systems. Recently my burst credit started to decline until it got to 0 so whenever the efs is used the throughput utilization reaches its maximum (100%). This makes the EFS SUPER slow and almost unreachable. The solution for the issue was to change the throughput mode from burst to elastic. So although the immediate problem was solved we still need to know which of the consumers that are mounted to this EFS consumes this much throughput, but I couldn't find any relevant metrics that show throughput consumption according to clients. I actually didn't find any metrics that show who the users of the EFS are.

I spent a lot of time trying to seek for a solution to see what is the consumption of each consumer, but really couldn't find an answer. Does anyone have an idea of how can I learn about the consumption of the clients mounted to this EFS?

thank you so much!

user15937765
  • 219
  • 1
  • 7

1 Answers1

1

AWS doesn't provide any metrics or logs or anything like what you are looking for with EFS. It's extremely difficult to deal with an issue like you describe. You will need to connect to each individual server that is using EFS and use the operating system's file system monitoring tools to determine how much each server is using the EFS volume. Of course if you are using AWS Lambda with the EFS volumes that won't be possible.

Mark B
  • 183,023
  • 24
  • 297
  • 295
  • Thank you for your answer! Do you know what is the best way to determine how much each server is using the EFS volume? For example, one of the server, runs a script every two minutes that go though all the folders in the EFS, but it only reads from the EFS, doesn't write, so I am not sure this server is my problem. Do you know a straight forward way to check what is the consumption each of the servers had (assuming I go inside each of the servers)? Thanks again – user15937765 Apr 25 '23 at 17:46
  • If they are Linux servers I would look into using `iotop`. – Mark B Apr 25 '23 at 19:08