Only process server updates on workloads affected by the server #12017
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
Signed-off-by: Alex Leong <[email protected]>
Signed-off-by: Alex Leong <[email protected]>
Signed-off-by: Alex Leong <[email protected]>
zaharidichev
approved these changes
Feb 1, 2024
|
Does this address #11995? |
|
@alpeb no it doesn't |
alpeb
added a commit
that referenced
this pull request
Feb 1, 2024
This edge release contains performance and stability improvements to the Destination controller, and continues stabilizing support for ExternalWorkloads. * Reduced the load on the Destination controller by only processing Server updates on workloads affected by the Server ([#12017]) * Changed how the Destination controller reacts to target clusters (in multicluster pod-to-pod mode) whose Server CRD is outdated: skip them and log an error instead of panicking ([#12008]) * Improved the leader election of the ExternalWorkloads Endpoints controller to avoid missing events ([#12021]) * Improved naming of EndpointSlices generated by ExternWorkloads ([#12016])
alpeb
added a commit
that referenced
this pull request
Feb 1, 2024
This edge release contains performance and stability improvements to the Destination controller, and continues stabilizing support for ExternalWorkloads. * Reduced the load on the Destination controller by only processing Server updates on workloads affected by the Server ([#12017]) * Changed how the Destination controller reacts to target clusters (in multicluster pod-to-pod mode) whose Server CRD is outdated: skip them and log an error instead of panicking ([#12008]) * Improved the leader election of the ExternalWorkloads Endpoints controller to avoid missing events ([#12021]) * Improved naming of EndpointSlices generated by ExternWorkloads ([#12016])
alpeb
added a commit
that referenced
this pull request
Feb 2, 2024
This edge release contains performance and stability improvements to the Destination controller, and continues stabilizing support for ExternalWorkloads. * Reduced the load on the Destination controller by only processing Server updates on workloads affected by the Server ([#12017]) * Changed how the Destination controller reacts to target clusters (in multicluster pod-to-pod mode) whose Server CRD is outdated: skip them and log an error instead of panicking ([#12008]) * Improved the leader election of the ExternalWorkloads Endpoints controller to avoid missing events ([#12021]) * Improved naming of EndpointSlices generated by ExternWorkloads ([#12016]) * Restriced the number of IPs an ExternalWorkload can have ([#12026])
alpeb
added a commit
that referenced
this pull request
Feb 2, 2024
This edge release contains performance and stability improvements to the Destination controller, and continues stabilizing support for ExternalWorkloads. * Reduced the load on the Destination controller by only processing Server updates on workloads affected by the Server ([#12017]) * Changed how the Destination controller reacts to target clusters (in multicluster pod-to-pod mode) whose Server CRD is outdated: skip them and log an error instead of panicking ([#12008]) * Improved the leader election of the ExternalWorkloads Endpoints controller to avoid missing events ([#12021]) * Improved naming of EndpointSlices generated by ExternWorkloads ([#12016]) * Restriced the number of IPs an ExternalWorkload can have ([#12026])
adleong
added a commit
that referenced
this pull request
Feb 19, 2024
We the destination controller's workload watcher receives an update for any Server resource, it recomputes opaqueness for every workload. This is because the Server update may have changed opaqueness for that workload. However, this is very CPU intensive for the destination controller, especially during resyncs when we get Server updates for every Server resource in the cluster. Instead, we only need to recompute opaqueness for workloads that are selected by the old version of the Server or by the new version of the Server. If a workload is not selected by either the new or old version of the Server, then the Server update cannot have changed the workload's opaqueness. Signed-off-by: Alex Leong <[email protected]>
Merged
adleong
added a commit
that referenced
this pull request
Feb 20, 2024
This stable release back-ports bugfixes and improvements from recent edge releases. * Introduced support for arbitrary labels in the `podMonitors` field in the control plane Helm chart (thanks @jseiser!) ([#11222]; fixes [#11175]) * Added a `prometheusUrl` field for the heartbeat job in the control plane Helm chart (thanks @david972!) ([#11343]; fixes [#11342]) * Updated the Destination controller to return `INVALID_ARGUMENT` status codes properly when a `ServiceProfile` is requested for a service that does not exist. ([#11980]) * Reduced the load on the Destination controller by only processing Server updates on workloads affected by the Server ([#12017]) * Changed how updates to a `Server` selector are handled in the destination service. When a `Server` that marks a port as opaque no longer selects a resource, the resource's opaqueness will reverted to default settings ([#12031]; fixes [#11995]) * Fixed a race condition in the destination service that could cause panics under very specific conditions ([#12022]; fixes [#12010]) * Fixed an issue where inbound policy could be incorrect after certain policy resources are deleted ([#12088]) [#11222]: #11222 [#11175]: #11175 [#11343]: #11343 [#11342]: #11342 [#11980]: #11980 [#12017]: #12017 [#11995]: #11995 [#12031]: #12031 [#12010]: #12010 [#12022]: #12022 [#12088]: #12088 Signed-off-by: Alex Leong <[email protected]> Signed-off-by: David ALEXANDRE <[email protected]> Signed-off-by: Justin S <[email protected]> Co-authored-by: Oliver Gould <[email protected]> Co-authored-by: Alejandro Pedraza <[email protected]> Co-authored-by: David ALEXANDRE <[email protected]> Co-authored-by: Justin Seiser <[email protected]>
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.
We the destination controller's workload watcher receives an update for any Server resource, it recomputes opaqueness for every workload. This is because the Server update may have changed opaqueness for that workload. However, this is very CPU intensive for the destination controller, especially during resyncs when we get Server updates for every Server resource in the cluster.
Instead, we only need to recompute opaqueness for workloads that are selected by the old version of the Server or by the new version of the Server. If a workload is not selected by either the new or old version of the Server, then the Server update cannot have changed the workload's opaqueness.