Performance tuning help

Elastic Observability APM
1

Using an Elastic cloud instance with 3 x APM 512 MB instances, but no matter what config performance tuning I do, I cannot get the "queue is full" spikes to stop.

Current configuration is as follows:

output.elasticsearch.worker: 32
output.elasticsearch.bulk_max_size: 5000
queue.mem.events: 50000

bulk_max_size and mem.events are both set to the max that Elastic cloud will allow and looking at the memory usage chart, we still have plenty unused. I have attempted many different combinations for these values with no success. Next looking at the elastisearch instances they are showing fine for disk usage, index speed, and cpu usage.

image
image.png2330×1191 340 KB

APM version 7.4.2

2

I was able to solve the issue by scaling up to 3 x APM 2 GB in elastic cloud. But after scale up the cpu usage and memory still showed the same in Stack Monitoring as show in my original post.

3

This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.