🌱 Add debug logging for the state of the priority queue #3075
Conversation
This debug logging prints the state of the workqueue in order to allow
debugging it. It will continously check at runtime if debug logging is
enabled and do nothing if not, making it very cheap if unused.
Sample output piped through `jq` for readability:
```
{
"level": "debug",
"ts": "2025-01-19T12:00:43-05:00",
"msg": "workqueue_state",
"controller": "configmap",
"items": [
{
"key": {
"Namespace": "kube-system",
"Name": "kubeadm-config"
},
"addedCounter": 1,
"priority": -100
}
]
}
```
k8s-ci-robot
added
the
cncf-cla: yes
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: alvaroaleman 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 |
k8s-ci-robot
added
approved
| addedCounter uint64 | ||
| priority int | ||
| readyAt *time.Time | ||
| Key T `json:"key"` |
There was a problem hiding this comment.
Exporting these is required for the json marshalling, otherwise they get omitted
| // Log level may change at runtime, so keep the | ||
| // loop going even if a given level is currently | ||
| // not enabled. | ||
| if !w.log.V(1).Enabled() { |
There was a problem hiding this comment.
I know its a bit more common in the ecosystem to use V(4) for debug, but I think many of our users use zap + zapr as the underlying implementation for the logr.Logger, given that we include it and do that in our examples. The problem is that zap has named log levels. These do end up mapping to a int8 but I personally only learned this recently and I've been using zap for years. The named debug level maps to -1 and zapr then switches the sign, resulting in V(1) in the logr.Logger.
Unless we have a strong reason why we need something higher than V(1), i'd like to keep it this way so that this log shows up if someone uses a zap logger on debug level.
This debug logging prints the state of the workqueue in order to allow debugging it. It will continously check at runtime if debug logging is enabled and do nothing if not, making it very cheap if unused.
Sample output piped through
jqfor readability: