-
Notifications
You must be signed in to change notification settings - Fork 33
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
Release 1.0.0-beta4 #26
Comments
I've been doing some thinking and I realised that the hostname label which you helped introduce in #25 will only return the hostname of the Pod (if run in a container). So my plan is to also add a Kubernetes specific label which will return the name of the node. I believe that was the desired behaviour from the beginning. |
Hi @amimof, Yeah, exactly today I bumped into the same issue. It seems to only work as expected if deployed as a native service and not as a container. I was able to get desired behaviour with Prometheus Please, let me know if I can assist in testing, implementation or something. 😄 |
@burningalchemist Yes we have also been using a |
I've created a PR implementing the expected behaviour for retrieving the Nodename. Please let me know if looks alright. :) |
@alexmirkhaydarov it looks fine, but it also makes us unable to use it outside of Kubernetes, thus making Currently, Kubernetes nodeName label can be added with Prometheus, or correctly retrieved if It's up to @amimof, but I would not consider removing hostname completely, but rather having it as a separate label, or actually implementing a fallback to os.Hostname if NODE_NAME can't be found. I'd try the latter. |
@alexmirkhaydarov Yes correct. I don't believe that the hostname label is to be replaced. It has its uses as well as the node name. I haven't had this confirmed, but doesn't Kubernetes add default environment variables to pods such as the node name? |
@amimof Unfortunately not. All the environment variables get injected based upon data available in manifests. Some of the fields are mandatory for manifests, so it makes them look default. |
@burningalchemist Yes you're right. I've picked up at some-point it won't be useable outside of Kubernetes and so I was working on a fallback fix. But I think having as a separate label makes total sense :) @amimof |
@burningalchemist |
@alexmirkhaydarov Thanks. I didn't meant for you to remove the Downward API config in the deploy manifest. What I meant was that the admin is in control of their config. But may use the provided one from this repo which contains necessary config to include node name as an env var. Sorry for the misunderstanding. |
The removal of the downward API back has been reverted on the deploy manifest. This issue can be closed now as the PR has been merged with the necessary changes. |
Just released v1.0.0-beta.4. Thank you so much for your time and effort 👏 |
Hi @amimof ,
If you have time, could you please add
1.0.0-beta4
tag, as we have hostname in metrics now and also a bugfix for a certificate version representation since1.0.0-beta3
.Kind regards,
Sergei
The text was updated successfully, but these errors were encountered: