Add a way to change CheckInterval/UpdateInterval/HealthyChecks at runtime

Hello,

first, thanks a lot for the runner. We are trying it for our own GitLab CI infrastructure.

One problem we have is the aggressive check interval of 3 seconds. Every 3 seconds, we have a new connection to the CI. This cause some troubles with our throttling limits on the CI server.

In our case, we think that we can change this 3 seconds timeout to something higher, like 30 seconds or even 1 minute. This will help us in the deployment of the runners.

Is it possible to add some configurations for some of the values in common/consts.go? The most importants are I think CheckInterval, UpdateInterval and HealthyChecks. (and maybe MaxTraceOutputSize, but it's not related to this issue.

Thanks in advance for your answer.

-- Olivier

It's good idea to have it configurable globally. I'll be happy to accept such merge request.

Added improvement label

Hello,

I have created a merge request which add a way to setting:

• CheckInterval,
• UndateInterval,
• MaxTraceOutputSize.

Note: this is my first try in go...

Thanks for any comments.

-- Olivier

Milestone changed to v0.4.0

@PuckCh thanks for your support. I would prefer not to add additional configuration options. We will think about the way to make the communication with Coordinator more efficient.

Status changed to closed

Hello,

Any news for any way to configure check/update interval ? 3 seconds are very aggressive... and since CI is now included in Gitlab, the production.log is growing very fast.

There is 3 way to solve it :

• reconsider the merge request from this issue
• add more option in gitlab for log/not log check requests
• review the communication beetween runner/gitlab

Thx in advance for any update :)

review the communication beetween runner/gitlab

I'm looking into this as proper resolution for this issue. It should happen in one of the upcoming releases.

Currently the Runner <-> CI communication is highly inefficient.