Using the same Elastic APM integration policies on 2 Fleet Servers

Kibana version: 8.14.0

Elasticsearch version: 8.14.0

APM Server version: 8.14.0

APM Agent language and version: Java Agent 1.38.0

I have installed Fleet server and I am using Apm agent integration. We monitor 98 applications. Using this as 2x Fleet Server and 1 elastic apm integration

Steps to reproduce:

Hello,

We monitor our Java applications through fleet server as Elastic Apm integration. There are almost 100 applications and this adds a lot of overhead. We catch most requests, but sometimes we get a request timed out error. I think the Fleet is having difficulty meeting the load amount. We have an idea to change the structure, use 2 fleet servers and use the same elastic APM integration to each of them. We think that this way we will distribute the load amount.

So can we add 1 elastic apm policies and use it on 2 fleet servers?

Is this possible? I expect your support regarding the error.

Fleet CPU usage: 200%

Thank you in advance for your support.

Errors in browser console (if relevant):

{"log.level":"error","@timestamp":"2024-06-14T16:47:18.064+0300","message":"request timed out","component":{"binary":"apm-server","dataset":"elastic_agent.apm_server","id":"apm-default","type":"apm"},"log":{"source":"apm-default"},"service.name":"apm-server","http.request.id":"11de60ae-ebc7-49d6-822b-b3f36c9d53d2","log.logger":"request","http.request.method":"POST","http.request.body.bytes":413261,"error.message":"request timed out","log.origin":{"file.line":59,"file.name":"middleware/log_middleware.go","function":"github.com/elastic/apm-server/internal/beater/api.apmMiddleware.LogMiddleware.func1.1"},"url.original":"/intake/v2/events","user_agent.original":"apm-agent-java/1.38.0 ()","event.duration":308858697642,"http.response.status_code":503,"source.address":"10.90.200.208","ecs.version":"1.6.0","ecs.version":"1.6.0"}

Any suggestions on this?

Hi beijing,

You can use the same APM integration with one Fleet-server policy, and enroll 2 fleet-server tot he same policy.
About the error it looks like a 503, perhapas the fleet-server didn't respond because of overload.

1 Like