Perform discovery API calls only as needed #39
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@dhaiducek and @JustinKuli I think this should help with the flaky tests. |
|
/hold the tests seem to have a lot of client throttling log messages so it may not work as I had hoped |
mprahl
force-pushed
the
more-efficient-discovery
branch
from
1d356e9
to
1d7d5b6
Compare
|
/unhold @JustinKuli and @dhaiducek I figured out a better approach. |
|
/hold for better error handling |
mprahl
force-pushed
the
more-efficient-discovery
branch
from
1d7d5b6
to
c931adb
Compare
|
/unhold |
|
@dhaiducek and @JustinKuli sorry for the noise. This is now ready for review. |
JustinKuli
reviewed
Mar 22, 2023
| } | ||
| } | ||
|
|
||
| return schema.GroupVersionResource{}, false, fmt.Errorf("%w: %s", errNoVersionedResource, gvk.String()) |
There was a problem hiding this comment.
It might be nice to slightly distinguish between this error case and the NotFound case earlier. I can imagine a debug session from logs where these two cases being identical could be frustrating.
This commit changes the discovery REST mapper that was used to be targeted to that specific resource version. This is to reduce load on the Kubernetes API as well as to reduce the chance of long delays due to queries per second (QPS) exceeding the default limit. Signed-off-by: mprahl <mprahl@users.noreply.github.com>
mprahl
force-pushed
the
more-efficient-discovery
branch
from
65c055d
to
8f6a415
Compare
mprahl
force-pushed
the
more-efficient-discovery
branch
from
27f97d5
to
22a4613
Compare
This should address the flakiness. Signed-off-by: mprahl <mprahl@users.noreply.github.com>
Prior to this, dynamic watch library was instructed to watch constraints even when the corresponding ConstraintTemplate hadn't had a CRD generated yet. This caused the dynamic watch library to constantly refresh its discovery metadata, which is taxing to the Kubernetes API and can cause client throttling. Signed-off-by: mprahl <mprahl@users.noreply.github.com>
This will give us more control of how the events are sent and perhaps reduce flakiness. Signed-off-by: mprahl <mprahl@users.noreply.github.com>
mprahl
force-pushed
the
more-efficient-discovery
branch
from
22a4613
to
bd19f2d
Compare
mprahl
added a commit
to mprahl/kubernetes-dependency-watches
that referenced
this pull request
Mar 23, 2023
Generating a REST mapper is very expensive on the Kubernetes API. This approach just gets the information needed for the particular watch request and then caches the information. This relates to this PR: open-cluster-management-io/governance-policy-framework-addon#39 Signed-off-by: mprahl <mprahl@users.noreply.github.com>
|
@dhaiducek and @JustinKuli note that if we merge this related PR, case12 would be much faster since it wouldn't cause the discovery information to be refreshed every time we try to create a watch on a Kind that does not exist: |
JustinKuli
approved these changes
Mar 23, 2023
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: JustinKuli, mprahl 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 |
mprahl
added a commit
to mprahl/kubernetes-dependency-watches
that referenced
this pull request
Mar 27, 2023
Generating a REST mapper is very expensive on the Kubernetes API. This approach just gets the information needed for the particular watch request and then caches the information. This relates to this PR: open-cluster-management-io/governance-policy-framework-addon#39 Signed-off-by: mprahl <mprahl@users.noreply.github.com>
openshift-merge-robot
pushed a commit
to stolostron/kubernetes-dependency-watches
that referenced
this pull request
Mar 27, 2023
Generating a REST mapper is very expensive on the Kubernetes API. This approach just gets the information needed for the particular watch request and then caches the information. This relates to this PR: open-cluster-management-io/governance-policy-framework-addon#39 Signed-off-by: mprahl <mprahl@users.noreply.github.com>
mprahl
pushed a commit
to mprahl/governance-policy-framework-addon
that referenced
this pull request
Apr 19, 2023
…cy (open-cluster-management-io#39) Adds a label to policy templates that are created as a child of a policy, then compares the policy templates actually contained in that policy to all templates with the parent policy label to determine if templates have been renamed or removed from the parent. The template sync will then clean up any policy templates that are not being managed by the parent policy that created them. refs: - https://issues.redhat.com/browse/ACM-3049?filter=-1 Signed-off-by: Will Kutler <wkutler@redhat.com> (cherry picked from commit 2661cb2) Co-authored-by: Will Kutler <wkutler@redhat.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This commit changes the discovery REST mapper that was used to be targeted to
that specific resource version.
This is to reduce load on the Kubernetes API as well as to reduce the chance
of long delays due to queries per second (QPS) exceeding the default limit.
An additional commit was added to prevent watches being created on Gatekeeper constraints without the corresponding CRD installed.
Other commits are to address flaky tests.
Relates:
https://issues.redhat.com/browse/ACM-4518