-
Notifications
You must be signed in to change notification settings - Fork 935
add manifests for initial deployment of sippy #1615
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
Conversation
groups/groups.yaml
Outdated
| - k8s-infra-rbac-triageparty-release@kubernetes.io | ||
| - k8s-infra-rbac-slack-infra@kubernetes.io | ||
| - k8s-infra-rbac-node-perf-dash@kubernetes.io | ||
| - k8s-infra-rbac-wg-reliability@kubernetes.io |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: This is already merged - please rebase
| wg-reliability has tools used to aid in transparency of stability metrics. | ||
|
|
||
| [sippy](https://github.com/openshift/sippy) summarizes multiple test-grid dashboards with data slicing of related jobs, job runs, | ||
| and tests. Visit (placeholder for eventual URL) to see the kube instance. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We should probably request public IP and DNS now (given we already have support for running it).
I'm wondering if we want "sippy.k8s.io" or something more generic?
| @@ -0,0 +1,17 @@ | |||
| apiVersion: v1 | |||
| kind: Service | |||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We also want Ingress and Certificate too, as e.g. here:
https://github.com/kubernetes/k8s.io/tree/master/perf-dash.k8s.io
| - --local-data | ||
| - /data | ||
| - --dashboard | ||
| - kube-master=sig-release-master-blocking,sig-release-master-informing= |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we want both? Or maybe we want to start just with blocking?
Or actually, I guess it's probably moot anyway, because we probably want to create our own dashboard for this purpose...
| terminationMessagePath: /dev/termination-log | ||
| terminationMessagePolicy: File | ||
| command: | ||
| - /tmp/src/scripts/fetchdata-kube.sh |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What is this? I didn't see that anywhere.
I think you can simply run:
./sippy --fetch-data /usr/local/google/home/wojtekt/sippy_data --dashboard=kube-master=sig-release-master-blocking=
Or I guess, you might wanted this to run periodically already.
Do we want a container? Or do we want a CronJob that will be doing this?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ahh - I guess with CronJob, we may have a problem with mounting for different pods...
So we probably want this to be a container within a pod.
| - mountPath: /data | ||
| name: data | ||
| restartPolicy: Always | ||
| terminationGracePeriodSeconds: 30 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Isn't 30 default?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it is
wg-reliability/sippy/pvc.yaml
Outdated
| - ReadWriteOnce | ||
| resources: | ||
| requests: | ||
| storage: 900M |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I guess we want to set some StorageClass, right?
|
/retest |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would rather see these manifests in a folder called sippy or sippy.k8s.io, in a sippy namespace to match.
If we're going down the path of assigning group ownership to apps deployed in cluster (we should! accountability is great!) I would rather see a sig: foo label, given that wg's don't own code.
I would suggest sig: architecture or sig: release to start... though long term it should be sig: testing when we have bandwidth.
|
/hold |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This looks fine to me (modulo missing storage class).
sippy/sippy/pvc.yaml
Outdated
| resources: | ||
| requests: | ||
| storage: 900M | ||
| # storageClassName: standard |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@spiffxp - what storage class we have in the community cluster?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@wojtek-t we have 2 storage classes in the cluster :
kubectl get sc
NAME PROVISIONER AGE
ssd kubernetes.io/gce-pd 334d
standard (default) kubernetes.io/gce-pd 551dThere was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks!
@deads2k - standard should be good enough for us;
Though I would visibly increase the size, say to 10GB or sth.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@deads2k - friendly ping
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
/approve
/lgtm
/hold
Remove /hold when ready to deploy
Thanks for changing the namespace, I ended up setting it up via #1844 so you'll likely need to rebase
|
|
||
| - email-id: k8s-infra-rbac-wg-reliability@kubernetes.io | ||
| name: k8s-infra-rbac-wg-reliability | ||
| - email-id: k8s-infra-rbac-sippy@kubernetes.io |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You'll probably need to rebase, I did this in #1844
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
/lgtm
sippy/sippy/ingress.yaml
Outdated
| @@ -0,0 +1,17 @@ | |||
| apiVersion: networking.k8s.io/v1 | |||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not sure if aaa is at a version that supports this
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
good catch - the cluster is still in 1.16 (will be upgraded to 1.17 soon, but ingress/v1 is only in 1.19), so we should use v1beta1 stil...
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: deads2k, spiffxp The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
|
/lgtm |
|
/hold We will merge on monday when everyone is back in the office and ready for potential excitement |
|
/hold cancel here we go! |
|
@deads2k for even more excitement. Is there a specific issue where I can follow the implementation? |
sippy lives here: https://github.com/openshift/sippy/ |
|
Ah sorry that was misleading. I meant more like the deployment of sippy for the kubernetes repos. I saw the DNS name |
As it happens, I was just starting to look now and I haven't worked out where my resources went. |
|
Ah got it, thx :). I'll just follow the update here then. |
|
@deads2k 👋🏾 , resources deployment is still a manual action in |
|
Part of #1900 |
builds on #1614
This adds manifests to deploy and run sippy. This is similar to what I have in my private toy cluster, but I need to do a little bit of alignment and work out ingresses instead of routes.
/assign @wojtek-t
/hold
explicit hold until I'm able to test these locally.