APM - Unexpected timeouts occurring

[Mike_H (Mike Hardy)]

Hi There, big fans of APM.

However, we are now getting timeout errors between the python agent and the apm server. We're on Python 3.6.6, Django 2.1.2 and Elasticsearch/Kibana/APM 6.4.1.

Previously, APM worked fine, but we've upgraded both Django and APM, and now timeouts are being thrown.

When runningpython manage.py elasticapm check, I get success messages.

When running python manage.py elasticapm test, I get PID 29: ElasticAPM is attempting to send 1 pending messages

When running the APM tutorial from the kibana dashboard, that is successful.

However, the APM dashboard shows the message, "No services found".

We're seeing regular errors like: [elasticapm.errors] [ERROR] Connection to APM Server timed out (url: http://elastic-apm.default.svc.cluster.local:8200/v1/transactions, timeout: None seconds)

All of this is deployed on kubernetes cluster, and we previously had this working by following the documented setup, so I'm not sure what we're missing.

Here's the relevant section from our settings.py

ELASTIC_APM = {
  'SERVICE_NAME': 'elastic-apm',
  'SERVER_URL': 'http://elastic-apm.default.svc.cluster.local:8200',
  'SERVER_TIMEOUT': 60,
  'DEBUG': True,
}

apm-server.host has a value of 0.0.0.0:8200

Help is much appreciated!

[Mike_H (Mike Hardy)]

Turns out it was an issue with our kubernetes cluster, which is still weird since we don't recall changing anything other than package versions, but nonetheless, it appears to be something other than APM.