-
Notifications
You must be signed in to change notification settings - Fork 718
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
Remove kube-rbac-proxy #8302
Remove kube-rbac-proxy #8302
Conversation
I'll look at this today @pebrc |
When testing with the following values config:
metrics:
port: 8443
secureMode:
enabled: true
image:
repository: docker.elastic.co/eck-ci/eck-operator-pr
tag: 8302-b2670646
installCRDs: true
managedNamespaces: []
webhook:
enabled: true I see this working without issues: I'll have to test with more scenarios tomorrow and update. |
From my testing the option for providing your own certificate/key/ca is broken, as it was configured in the rbac-proxy in the previous implementation. I'm working through what the fix looks like for this and will update. |
Signed-off-by: Michael Montgomery <mmontg1@gmail.com>
I'll take a look at this and update... thanks |
This reverts commit 42e0d56.
i sat down to finish this and saw you handled it. i will verify in the am. thanks |
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.
Nice work! I really like the additional fixes/improvements you included in this PR 👍
I ran a couple of tests, only with Helm, using both the ECK and the Prometheus operator charts, with a custom CA, enforcing (or not) TLS validation, and it worked as expected.
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 good 👍
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.
Not from your PR, I noticed the metadata.labels
of the Service
and the spec.selector.matchLabels
of the ServiceMonitor
in configure-operator-metrics.asciidoc
are not aligned with what is rendered using the Helm chart.
This removes the kube-auth-proxy from the ECK Helm charts and the ECK documentation. Instead if follows the recommendation from controller-runtime to use the built-in FilterProvider filters.WithAuthenticationAndAuthorization. This pulls in a bunch of k8s API server dependencies increasing the binary size by about 12MB IIRC. I have also tried to address some issues with our current Helm templating of the metrics server: - allow enabling the secure mode with TLS+auth while not forcing users to have Promotheus installed e.g. when using Elastic Agent (!) (service monitor is still generated by default for bwc) - mixing configuration properties for the service monitor with the configuration properties of the metrics server (I moved them to serviceMonitor.* while implementing a form of bwc layer in the template --------- Co-authored-by: Michael Montgomery <mmontg1@gmail.com> Co-authored-by: Michael Morello <michael.morello@elastic.co> (cherry picked from commit 032bff5) # Conflicts: # go.sum
💚 All backports created successfully
Questions ?Please refer to the Backport tool documentation |
This removes the kube-auth-proxy from the ECK Helm charts and the ECK documentation. Instead if follows the recommendation from controller-runtime to use the built-in FilterProvider filters.WithAuthenticationAndAuthorization. This pulls in a bunch of k8s API server dependencies increasing the binary size by about 12MB IIRC. I have also tried to address some issues with our current Helm templating of the metrics server: - allow enabling the secure mode with TLS+auth while not forcing users to have Promotheus installed e.g. when using Elastic Agent (!) (service monitor is still generated by default for bwc) - mixing configuration properties for the service monitor with the configuration properties of the metrics server (I moved them to serviceMonitor.* while implementing a form of bwc layer in the template --------- Co-authored-by: Michael Montgomery <mmontg1@gmail.com> Co-authored-by: Michael Morello <michael.morello@elastic.co> (cherry picked from commit 032bff5)
Fixes #8279
This removes the kube-auth-proxy from the ECK Helm charts and the ECK documentation. Instead if follows the recommendation from controller-runtime to use the built-in
FilterProvider
filters.WithAuthenticationAndAuthorization
.This pulls in a bunch of k8s API server dependencies increasing the binary size by about 12MB IIRC.
I have also tried to address some issues with our current Helm templating of the metrics server:
serviceMonitor.*
while implementing a form of bwc layer in the templateThings I did not address:
Tests I did (all with Prometheus operator):
insecureSkipVerify: true
insecureSkipVerify: true
insecureSkipVerify: false
Review
If you can test a few scenarios while reviewing this it would much appreciated to flush out any issues I have missed.