-
Notifications
You must be signed in to change notification settings - Fork 168
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[alertmanager] Add config to disable clustering #460
[alertmanager] Add config to disable clustering #460
Conversation
Thanks @ionphractal for your PR. The alertmanager documentation states that empty string should be given to Could you rework this PR in order to stick with this approach ? Remove the introduced Also, the tip given by sylr in prometheus/alertmanager#1814 looks suspicious. Given the following code in alertmanager: There should be no difference between @benjaminguttmann-avtq any thoughts on this ? |
…tly as an empty string
Technically it should not make any difference because of the equal "=" sign followed by a space, but for clarity I added the double quotes. I tested it in our system that it works that way as well. Regarding the |
@ionphractal normally, I would agree with you. I 100% prefer a clear disable option than relying on a particular value that encodes this behaviour. However, bosh releases are not software per say but only a packaging system. In this particular case, I prefer transparency and stick with design decisions made by underlying software when possible (even if I consider these as "poor" decisions). Since both approaches has pros and cons, lets settle this by asking @benjaminguttmann-avtq his opinion and go in his direction. |
I can understand both arguments but I would prefer the clear 'enabled' statement then having an empty string, because that could in theory happen by accident whereas the other one is a clear decision. |
so be it, lets go for |
Here you go :) |
thanks @ionphractal |
When only one alertmanager instance is deployed, alertmanager_cluster_messages_queued will rise over time until it hits 4k queue limit resulting in dropping alerts (prometheus/alertmanager#1814). This PR enables operators to disable clustering for alertmanager.