⚠️ Allow configuring more granular cache filtering #2421
Conversation
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
the
approved
k8s-ci-robot
added
the
size/XXL
alvaroaleman
changed the title
alvaroaleman
changed the title
alvaroaleman
force-pushed
the
cache
branch
6 times, most recently
from
31e52e1
to
586bca7
Compare
pkg/cache/delegate_by_type.go
Outdated
| synced := true | ||
| for _, cache := range append(maps.Values(dbt.caches), dbt.defaultCache) { | ||
| if !cache.WaitForCacheSync(ctx) { | ||
| synced = false |
There was a problem hiding this comment.
| synced = false | |
| return false |
Q: Would this make sense here?
If yes, we have a few more places where we could do the same (but maybe better in a separate PR)
There was a problem hiding this comment.
The reason it is the way that it is is that the interface is documented as "Blocks until caches are synced". If we return early, not all caches might be synced. I don't know if anyone would continue their binary if this returns false but odds are, someone somewhere will
There was a problem hiding this comment.
Yeah let's better not change it. Was just thinking if we're hitting the case where some caches are definitely not synced we might as well return earlier. But no real reason to make a change
This change implements most of the [cache options design][0], in particular it is now possible to: * Configure Namespaces per type * Configure filtering per namespace What is still missing is to allow configuring namespace-level default settings that will be used for all namespaces not explicitly configured. There is nothing in the way of doing that as a follow-up. The implementation slightly derivates from the design document in that the filter settings in `ByGVK` do not use the `Config` struct but instead are directly embedded. This allows us to not break compatibility and is more ergonomic to use. The implementation is based on a new and internal per-type delegating "meta cache". [0]: https://github.com/kubernetes-sigs/controller-runtime/blob/main/designs/cache_options.md
|
@alvaroaleman: The following test failed, say
Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR. 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. I understand the commands that are listed here. |
|
Thx! /lgtm |
k8s-ci-robot
added
the
lgtm
This change implements most of the cache options design, in particular it is now possible to:
What is still missing is to allow configuring namespace-level default settings that will be used for all namespaces not explicitly configured. There is nothing in the way of doing that as a follow-up.
The implementation slightly derivates from the design document in that the filter settings in
ByGVKdo not use theConfigstruct but instead are directly embedded. This allows us to not break compatibility and is more ergonomic to use.The implementation is based on a new and internal per-type delegating "meta cache".