Package org.apache.flink.kubernetes.operator.utils

Class StatusRecorder<CR extends AbstractFlinkResource<?,STATUS>,STATUS extends CommonStatus<?>>

java.lang.Object

org.apache.flink.kubernetes.operator.utils.StatusRecorder<CR,STATUS>

public class StatusRecorder<CR extends AbstractFlinkResource<?,STATUS>,STATUS extends CommonStatus<?>>
extends java.lang.Object

Helper class for status management and updates.

Field Summary

Fields

Modifier and Type	Field	Description
protected com.fasterxml.jackson.databind.ObjectMapper	objectMapper
protected java.util.concurrent.ConcurrentHashMap<io.javaoperatorsdk.operator.processing.event.ResourceID,com.fasterxml.jackson.databind.node.ObjectNode>	statusCache

Constructor Summary

Constructors

Constructor	Description
StatusRecorder(MetricManager<CR> metricManager, java.util.function.BiConsumer<CR,STATUS> statusUpdateListener)

Method Summary

Modifier and Type	Method	Description
static <S extends CommonStatus<?>,CR extends AbstractFlinkResource<?,S>> StatusRecorder<CR,S>	create(io.fabric8.kubernetes.client.KubernetesClient kubernetesClient, MetricManager<CR> metricManager, java.util.Collection<FlinkResourceListener> listeners)
void	patchAndCacheStatus(CR resource, io.fabric8.kubernetes.client.KubernetesClient client)	Update the status of the provided kubernetes resource on the k8s cluster.
void	removeCachedStatus(CR resource)	Remove cached status for Flink resource.
void	updateStatusFromCache(CR resource)	Update the custom resource status based on the in-memory cached to ensure that any status updates that we made previously are always visible in the reconciliation loop.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

objectMapper

protected final com.fasterxml.jackson.databind.ObjectMapper objectMapper

statusCache

protected final java.util.concurrent.ConcurrentHashMap<io.javaoperatorsdk.operator.processing.event.ResourceID,com.fasterxml.jackson.databind.node.ObjectNode> statusCache

Constructor Detail

StatusRecorder

public StatusRecorder(MetricManager<CR> metricManager,
                      java.util.function.BiConsumer<CR,STATUS> statusUpdateListener)

Method Detail

patchAndCacheStatus

public void patchAndCacheStatus(CR resource,
                                io.fabric8.kubernetes.client.KubernetesClient client)

Update the status of the provided kubernetes resource on the k8s cluster. We use patch together with null resourceVersion to try to guarantee that the status update succeeds even if the underlying resource spec was update in the meantime. This is necessary for the correct operator behavior.

Parameters:

resource - Resource for which status update should be performed

updateStatusFromCache

public void updateStatusFromCache(CR resource)

Update the custom resource status based on the in-memory cached to ensure that any status updates that we made previously are always visible in the reconciliation loop. This is required due to our custom status patching logic.

If the cache doesn't have a status stored, we do no update. This happens when the operator reconciles a resource for the first time after a restart.

Parameters:

resource - Resource for which the status should be updated from the cache

removeCachedStatus

public void removeCachedStatus(CR resource)

Remove cached status for Flink resource.

Parameters:

resource - Flink resource.

create

public static <S extends CommonStatus<?>,CR extends AbstractFlinkResource<?,S>> StatusRecorder<CR,S> create(io.fabric8.kubernetes.client.KubernetesClient kubernetesClient,
                                                                                                                              MetricManager<CR> metricManager,
                                                                                                                              java.util.Collection<FlinkResourceListener> listeners)