Skip to content
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

⚠️ Fix broken kind logging #1954

Merged
merged 2 commits into from
Jul 15, 2022

Conversation

tatsuhiro-t
Copy link
Contributor

@tatsuhiro-t tatsuhiro-t commented Jul 13, 2022

If gvk.Kind is "APIService", I expect "apiService", but the current
code produces and logs "aPIService". Because it is not easy to do
this kind of thing, it is better to log the kind as is.

Fixes #1945


This is a breaking change. Previously, the log line included a reconciled resource (for example, APIService bar/foo) like:

"aPIService"={"name":"foo","namespace":"bar"}

Now it looks like:

"APIService"={"name":"foo","namespace":"bar"}

If gvk.Kind is "APIService", I expect "apiService", but the current
code produces and logs "aPIService".  Because it is not easy to do
this kind of thing, it is better to log the kind as is.
@k8s-ci-robot k8s-ci-robot added cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels Jul 13, 2022
@k8s-ci-robot
Copy link
Contributor

Hi @tatsuhiro-t. Thanks for your PR.

I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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.

@k8s-ci-robot k8s-ci-robot added the size/XS Denotes a PR that changes 0-9 lines, ignoring generated files. label Jul 13, 2022

ctrlOptions.LogConstructor = func(req *reconcile.Request) logr.Logger {
log := log
if req != nil {
log = log.WithValues(
lowerCamelCaseKind, klog.KRef(req.Namespace, req.Name),
kind, klog.KRef(req.Namespace, req.Name),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So why not just use gvk.Kind?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed via bb36890

@FillZpp
Copy link
Contributor

FillZpp commented Jul 13, 2022

/ok-to-test

@k8s-ci-robot k8s-ci-robot added ok-to-test Indicates a non-member PR verified by an org member that is safe to test. and removed needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels Jul 13, 2022
@FillZpp
Copy link
Contributor

FillZpp commented Jul 13, 2022

/label tide/merge-method-squash

/lgtm

@k8s-ci-robot k8s-ci-robot added the tide/merge-method-squash Denotes a PR that should be squashed by tide when it merges. label Jul 13, 2022
@k8s-ci-robot k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Jul 13, 2022
@FillZpp
Copy link
Contributor

FillZpp commented Jul 13, 2022

/cc @alvaroaleman @sbueringer

ctrlOptions.LogConstructor = func(req *reconcile.Request) logr.Logger {
log := log
if req != nil {
log = log.WithValues(
lowerCamelCaseKind, klog.KRef(req.Namespace, req.Name),
gvk.Kind, klog.KRef(req.Namespace, req.Name),
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is a breaking change for anyone who filters their logs, please mark it as such. Please also describe how this now looks like rather than only how it does not look like anymore.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Changed the icon to ⚠️ to indicate this is a breaking change, and edited the PR description to include how it looks now.

Copy link
Member

@sbueringer sbueringer Jul 14, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I see your point that aPIService is definitely bad. On the other side this change is directly in conflict with the conventions used by k/k: https://github.com/kubernetes/community/blob/master/contributors/devel/sig-instrumentation/migration-to-structured-logging.md#name-arguments

Specifically:

Always use lowerCamelCase, for example use containerName and not container name or container_name.

I wonder if there is any alternative instead of using kind directly

Copy link
Member

@sbueringer sbueringer Jul 14, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking at the example of good keys I'm not sure if APIService should be logged as APIService or apiService (especially the CIDR and PVC examples). Maybe we can ask the logging folks in the #klog Slack channel.

If we assume the first, it could be good enough to just keep the current case when the second character is also uppercase. (I'm aware it's not a great solution, but I assume it's a good heuristic)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

My thoughts:

  1. Things like PVC is just the short name of PersistentVolumeClaim, which is different with the real kind APIService.
  2. IMHO, changing the first case to lower is a little weird. Kubernetes resources have kind like PersistentVolumeClaim and resource(plural) like persistentvolumeclaims. But things like persistentVolumeClaim are only used as golang variable names for they should have first case lower, why should we keep this format in log?

Copy link
Contributor Author

@tatsuhiro-t tatsuhiro-t Jul 14, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

My first impression of the linked documentation is it does not take into account the CRD names which can be arbitrarily chosen by third parties and is hard to convert to camelCase in some cases.

Alternatively, we may follow the statement under exception lists:

When Kubernetes object kind is unknown without runtime checking we should use object key. To provide information about kind we should add separate apiVersion and kind fields.

so use object as a key to log name and namespace, and log kind as is under kind key.

"object"={"name":"foo","namespace":"bar"} "kind"="APIService"

Copy link
Member

@sbueringer sbueringer Jul 14, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have no idea why they picked those conventions, but I agree it's definitely hard to do if we can't just hard-code the key.

Given that, I think i also prefer just using kind directly.

The problem with using object is that it makes it very very hard to correlate log entries via a log query across controllers.
E.g. the following example:

  • 1st controller reconciles Clusters: k/v pairs: "Cluster": <custer>
  • 2nd controller reconciles Machines: k/v pairs: "Machine": <machine>, "Cluster": <cluster>
  • ...

By using kind it's trivial to match logs across controllers. If the reconciled object has the "object" key it requires to calculate the actual kind before correlation between logs of different controllers are possible.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 to use kind without transformations, it is simpler, less verbose than object and it read more naturally

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right, so if I understand it correctly, we have a consensus on what we have now in this PR, that is log kind as a key without transformation.

@tatsuhiro-t tatsuhiro-t changed the title 🐛 Fix broken kind logging ⚠️ Fix broken kind logging Jul 13, 2022
@sbueringer
Copy link
Member

/lgtm

@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: alvaroaleman, tatsuhiro-t

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 /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@k8s-ci-robot k8s-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Jul 15, 2022
@k8s-ci-robot k8s-ci-robot merged commit b93b5f9 into kubernetes-sigs:master Jul 15, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
approved Indicates a PR has been approved by an approver from all required OWNERS files. cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. lgtm "Looks good to me", indicates that a PR is ready to be merged. ok-to-test Indicates a non-member PR verified by an org member that is safe to test. size/XS Denotes a PR that changes 0-9 lines, ignoring generated files. tide/merge-method-squash Denotes a PR that should be squashed by tide when it merges.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Broken resource kind logging
6 participants