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
Bug 1809665: Tune AWS load balancers to be consistent with other platforms #363
Conversation
@sttts when we picked numbers for kube-apiserver, what were our bounds? Any considerations I might be missing? |
/retest |
1 similar comment
/retest |
4e276cd
to
c7c9a8b
Compare
GCP and Azure both hardcode LB timeouts - set AWS to match the lower bound of Azure (2 failures @ 5s interval, 2 healthy). By default AWS is 6 failures @ 10s which is a very long timeout for the router. In the future, if these attributes become customizable on GCP and Azure, they should be consistent with our AWS settings.
/retest |
We started out with 10 sec interval and threshold 3. We never questioned those numbers. The 70 sec minimal termination delay was just derived from that, working well in practice. So no strong analysis behind all of that. |
/assign @ironcladlou |
@@ -40,6 +40,7 @@ var ( | |||
"service.beta.kubernetes.io/aws-load-balancer-internal": "0.0.0.0/0", | |||
}, | |||
configv1.AzurePlatformType: { | |||
// Azure load balancers are not customizable and are set to (2 fail @ 5s interval, 2 healthy) |
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.
Looks like a leftover that needs deleted
/approve |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: ironcladlou, smarterclayton 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 |
/retitle Bug 1809665: Tune AWS load balancers to be consistent with other platforms |
@smarterclayton: Bugzilla bug 1809665 is in an unrecognized state (ON_QA) and will not be moved to the MODIFIED state. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
@Miciah: new pull request created: #370 In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/cherry-pick release-4.3 |
@Miciah: new pull request created: #371 In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
GCP and Azure both hardcode LB timeouts - set AWS to match the lower bound of Azure (2 failures @ 5s interval, 2 healthy). By default AWS is 6 failures @ 10s which is a very long timeout for the router.
In the future, if these attributes become customizable on GCP and Azure, they should be consistent with our AWS settings.
If we see flakiness at this lower check interval, we may want to consider bumping to 3 successive failures. However, it's generally best to start at the low end and then tune up to be as responsive
as possible to rollouts.