[elastic-agent][container] increase limit for retries to Kibana

[simitt]

The container cmd retries requests to Kibana up to 30 times with 1 second sleep intervals.
Kibana can take several minutes to start up. Increase the retry counter/timeout accordingly, or make it configurable. This is important for scenarios where Kibana and the Elastic Agent get deployed at the same time, such as in ESS/ECE.

[elasticmachine]

Pinging @elastic/agent (Team:Agent)

[ph]

@michalpristas Can you take a look at this one?
@simitt I would consider this a bug, but the risk is to have a disconnected instance of Elastic Agent, right?

And should we make it retry infinitely?

[blakerouse]

We should not make it go infinitely by default. It would be better to see the container restarted in Kubernetes versus it always up even though its not connected.

I think adding ENV variables to control this would be best.

[simitt]

making the timeout configurable SGTM;

but the risk is to have a disconnected instance of Elastic Agent, right?

yes, but in some cases that's preferable to having the container being restarted (because the elastic-agent is in beta, but apm-server also started from this process is GA).

[ph]

Ok, let's make it configurable.

[ph]

@simitt The retry PR were merged, let's confirm and close this issue?

[ruflin]

Closing, please reopen if issue still exists.