#### Network
- EndpointSlices: If upgrading a cluster with EndpointSlices already enabled, any EndpointSlices that should be managed by the EndpointSlice controller should have a `endpointslice.kubernetes.io/managed-by` label set to `endpointslice-controller.k8s.io`.

#### Scheduling

v1.25.16](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager) | [amd64](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager-amd64), [arm](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager-arm), [arm64](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager-arm64), [ppc64le](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-co...

v1.23.17](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager) | [amd64](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager-amd64), [arm](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager-arm), [arm64](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-controller-manager-arm64), [ppc64le](https://console.cloud.google.com/gcr/images/k8s-artifacts-prod/us/kube-co...

    health check fails.


## Version 2.7.0

_2015-12-13_

 *  **Rewritten connection management.** Previously OkHttp's connection pool
    managed both idle and active connections for HTTP/2, but only idle
    connections for HTTP/1.x. With this update the connection pool manages both
    idle and active connections for everything. OkHttp now detects and warns on
    connections that were allocated but never released, and will enforce HTTP/2

		*GetUser("bulk_5", Config{Account: false, Pets: 0, Toys: 3, Company: true, Manager: false, Team: 1, Languages: 3, Friends: 1}),
		*GetUser("bulk_6", Config{Account: true, Pets: 4, Toys: 3, Company: false, Manager: true, Team: 1, Languages: 3, Friends: 0}),
		*GetUser("bulk_7", Config{Account: true, Pets: 1, Toys: 3, Company: true, Manager: true, Team: 4, Languages: 3, Friends: 1}),

     * @return the maximum cache age in milliseconds
     */
    public long getMaxAge() {
        return maxAge;
    }

    /**
     * Sets the maximum age for cached data
     * @param maxAge the maximum cache age in milliseconds
     */
    public void setMaxAge(long maxAge) {
        this.maxAge = maxAge;
    }

    /**
     * Gets the timestamp when this cache entry was created

                // Don't need to interrupt thread since we never start it in tests
            }
        };
        manager.init();
        assertTrue(manager.baseDir.exists());
        manager.destroy();
    }

    // Test initialization failure
    public void test_init_failure() {
        ThumbnailManager manager = new ThumbnailManager() {
            @Override
            public void init() {

### Async Context Manager { #async-context-manager }

If you check, the function is decorated with an `@asynccontextmanager`.

That converts the function into something called an "**async context manager**".

{* ../../docs_src/events/tutorial003.py hl[1,13] *}

    private final ChannelManager manager;
    private final ExecutorService executor;
    private final ConcurrentMap<String, FailoverState> failoverStates;

    /**
     * Create channel failover handler
     *
     * @param manager channel manager
     */
    public ChannelFailover(ChannelManager manager) {
        this.manager = manager;
        this.executor = Executors.newCachedThreadPool(r -> {

		*GetUser("delete_slice_with_associations2", Config{Account: true, Pets: 3, Toys: 2, Company: true, Manager: true, Team: 2, Languages: 2, Friends: 3}),
		*GetUser("delete_slice_with_associations3", Config{Account: true, Pets: 2, Toys: 3, Company: true, Manager: true, Team: 3, Languages: 3, Friends: 2}),