0

My system sends the data to Redis via web server and ultimately stacks it in Elasticsearch. The thing is that Redis connection timeout error occurs when the traffic is small. The error always occurs at 5 AM in Korean time around one to five minutes. The memory usage rate of the Redis server rapidly increases at 5 AM, but still it’s only about 2GB. Currently I’m using the Redis premium price tier for your reference. BTW, I believe that my process doesn’t include any scheduler that can cause the rapid increase of the memory usage rate at 5 AM. Is there any chance that the error came from the Redis or Elasticsearch? I don’t see any problematic log around the error occurred time in the Elasticsearch. I look forward to your help.

Best Regards, Snow

Gwi Hyeon Jeon
  • 1

1 Answers1

0

That could come from many issues :

Because you use the SAVE command (it is setup in your conf) generating a lot of I/O and hammering the server on that time or any backups are scheduled, especially if you use EBS volumes on Amazon. Also if you use a KEY * which is very slow on a lot of indexes. Try "slow log" command on the Redis server to see if there are some "slow query". Write some logs when "TimeoutError" happens, to see if the "error Redis command" in the "slow log" and adjust your timeout setting on the client-side.

Thanks,

Aiden
  • 177
  • 2