Use standard liveness and readiness settings as KC 22.0.5 is released

doc/kubernetes/modules/ROOT/pages/running/concepts/threads.adoc

@@ -56,14 +56,9 @@ Future version of Keycloak will have better means to handle that: https://github
 [#probes]
 === Probes
 
-All health probes, including liveness and readiness probes, are handled in the Quarkus executor worker pool.
-With all requests being queued, also the liveness probe is queued, and is therefore slow.
-During Keycloak Infinispan view updates for members leaving and rebalancing, there is an increased latency for all requests, observed with up to 10 seconds.
-In a high-load or even overload scenario, the probes will be queued in the executor thread pool, and won't return in time.
-When requests queue up in Keycloak, also Readiness and Liveness probes are delayed, which might trigger failure detection in Kubernetes and will lead to Pod restarts in overload or load-shedding situations.
-With load shedding activated, when requests are rejected from the executor thread pool, failing readiness probes will lead to Pods not receiving any load for a period or time, and with failing liveness probes the Pods will eventually be restarted.
-This is tracked in https://github.com/keycloak/keycloak/issues/22109[keycloak#22109].
-For the time being, consider a longer timeout for the probes to survive spikes in the delay, or disabling the liveness probe to avoid Pod restarts.
+All health probes, including liveness and readiness probes, are handled in the Quarkus executor worker pool by default.
+Starting with Keyloak 22.0.5, there is a workaround in place to have the liveness probe non-blocking (see: https://github.com/keycloak/keycloak/issues/22109[keycloak#22109]).
+Future version of Keycloak and Quarkus plan to have other probes also being non-blocking.
 
 === OS Resources
 

doc/kubernetes/modules/ROOT/pages/running/keycloak-deployment.adoc

@@ -42,9 +42,12 @@ See xref:./concepts/threads.adoc[] for details.
 <4> The JVM options set additional parameters:
 * `jgroups.thread_dumps_threshold` ensures that a log message "`thread pool is full`" appears once the JGroup thread pool is full for the first time. See xref:./concepts/threads.adoc[] for details.
 * Adjust the memory settings for the heap.
-<5> Probes will be delayed when requests queue up under load, and fail when a load is shed.
-So the best way to run Keycloak in Kubernetes would be to disable those probes, for now.
-See xref:./concepts/threads.adoc#probes[Probes] for details.
+
+[NOTE]
+====
+Previous versions of this documentation recommended to disable the liveness and readiness probes.
+With https://github.com/keycloak/keycloak/issues/22109[keycloak#22109] being available in Keycloak 22.0.5, this is no longer recommended.
+====
 
 == Optional: Load shedding
 

provision/minikube/keycloak/templates/keycloak.yaml

@@ -257,15 +257,15 @@ spec:
               limits:
                 {{ if .Values.cpuLimits }}cpu: "{{ .Values.cpuLimits }}"{{end}}
                 {{ if .Values.memoryLimitsMB }}memory: "{{ .Values.memoryLimitsMB }}M"{{end}}
-            readinessProbe: # <5>
-              exec:
-                command:
-                  - 'true'
-            livenessProbe: # <5>
-              exec:
-                command:
-                  - 'true'
             # end::keycloak[]
+            # readinessProbe:
+            #   exec:
+            #     command:
+            #       - 'true'
+            # livenessProbe:
+            #   exec:
+            #     command:
+            #       - 'true'
             # tag::keycloak-ispn[]
             volumeMounts:
 {{ if .Values.infinispan.customConfig }}