Skip disabling keep-alives when shutting down the webhook server #3295
Conversation
My theory is that when this happens the go runtime will close idle connections. If by chance the API server is making a request and re-using a connection (that is closing) it will receive an io.EOF. Thus introducing a flake. So we now remove this setting change and let all the connections close during the regular graceful shutdown. We would have already lame ducked so the webhook shouldn't be receiving new requests.
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: dprotaso The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
knative-prow
bot
added
the
approved
knative-prow
bot
added
the
size/XS
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #3295 +/- ##
==========================================
- Coverage 74.60% 74.58% -0.03%
==========================================
Files 188 188
Lines 8188 8187 -1
==========================================
- Hits 6109 6106 -3
- Misses 1840 1841 +1
- Partials 239 240 +1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
/assign @nader-ziada @linkvt |
|
/lgtm |
|
/cherry-pick release-1.20 |
|
@dprotaso: new pull request created: #3305 DetailsIn 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-sigs/prow repository. |
|
/cherry-pick release-1.19 |
|
@dprotaso: new pull request created: #3306 DetailsIn 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-sigs/prow repository. |
4 participants
Fixes: #1509
My theory is that when this happens the go runtime will close
idle connections. If by chance the API server is making a
request and re-using a connection (that is closing) it will receive
an io.EOF. Thus introducing a flake.
So we now remove this setting change and let all the connections
close during the regular graceful shutdown. We would have
already lame ducked so the webhook shouldn't be receiving new
requests.