docs: skip reconcile of a DR (#2732)

Signed-off-by: Attila Mészáros <[email protected]>
Signed-off-by: Chris Laprun <[email protected]>
Co-authored-by: Chris Laprun <[email protected]>

Author: csviri

docs/content/en/docs/faq/_index.md

@@ -3,7 +3,7 @@ title: FAQ
 weight: 80
 ---
 
-### Q: How can I access the events which triggered the Reconciliation?
+### How can I access the events which triggered the Reconciliation?
 
 In the v1.* version events were exposed to `Reconciler` (which was called `ResourceController`
 then). This included events (Create, Update) of the custom resource, but also events produced by
@@ -16,7 +16,7 @@ sound agreement between the developers that this is the way to go.
 Note that this is also consistent with Kubernetes 
 [level based](https://cloud.redhat.com/blog/kubernetes-operators-best-practices) reconciliation approach. 
 
-### Q: Can I re-schedule a reconciliation, possibly with a specific delay?
+### Can I re-schedule a reconciliation, possibly with a specific delay?
 
 Yes, this can be done
 using [`UpdateControl`](https://github.com/java-operator-sdk/java-operator-sdk/blob/main/operator-framework-core/src/main/java/io/javaoperatorsdk/operator/api/reconciler/UpdateControl.java)
@@ -46,7 +46,7 @@ without an update:
 Although you might consider using `EventSources`, to handle reconciliation triggering in a smarter
 way. 
 
-### Q: How can I run an operator without cluster scope rights?
+### How can I run an operator without cluster scope rights?
 
 By default, JOSDK requires access to CRs at cluster scope. You may not be granted such
 rights and you will see some error at startup that looks like:
@@ -74,7 +74,7 @@ is `true` (`false` by default). To disable, set it to `false` at [Operator-level
 Operator operator = new Operator( override -> override.checkingCRDAndValidateLocalModel(false));
 ```
 
-### Q: I'm managing an external resource that has a generated ID, where should I store that?
+### I'm managing an external resource that has a generated ID, where should I store that?
 
 It is common that a non-Kubernetes or external resource is managed from a controller. Those external resources might
 have a generated ID, so are not simply addressable based on the spec of a custom resources. Therefore, the 
@@ -91,8 +91,39 @@ it is not guaranteed that during the next reconciliation you will see the fresh
 which do this, usually cache the updated status in memory to make sure it is present for next reconciliation.
 
 Dependent Resources feature supports the [first approach](../dependent-resources/_index.md#external-state-tracking-dependent-resources).
+    
+### How can I skip the reconciliation of a dependent resource?
 
-### Q: How to fix `sun.security.provider.certpath.SunCertPathBuilderException` on Rancher Desktop and k3d/k3s Kubernetes
+Skipping workflow reconciliation altogether is possible with the explicit invocation feature since v5. 
+You can read more about this in [v5 release notes](https://javaoperatorsdk.io/blog/2025/01/06/version-5-released/#explicit-workflow-invocation).
+
+However, what if you want to avoid reconciling a single dependent resource based on some state?
+First of all, remember that the dependent resource won't be modified if the desired state and the actual state match.
+Moreover, it is generally a good practice to reconcile all your resources, JOSDK taking care of only processing the
+resources which state doesn't match the desired one.
+However, in some corner cases (for example, if it is expensive to compute the desired state or compare it to the actual
+state), it is somtimes useful to be able to only skip the reconcilation of some resources but not all, if it is known
+that they don't need to be processed based for example on the status of the custom resource.
+
+A common mistake is to use `ReconcilePrecondition`, if the condition does not hold it will delete the resources.
+This is by design (although it's true that the name of this condition might be misleading), but not what we want in this
+case.
+
+The way to go is to override the matcher in the dependent resource:
+
+```java
+public Result<R> match(R actualResource, R desired, P primary, Context<P> context) {
+    if (alreadyIsCertainState(primary.getStatus())) {
+        return true;
+    } else {
+        return super.match(actual, desired, primary, context);
+    }
+}
+```
+
+This will make sure that the dependent resource is not updated if the primary resource is in certain state.
+
+### How to fix `sun.security.provider.certpath.SunCertPathBuilderException` on Rancher Desktop and k3d/k3s Kubernetes
 
 It's a common issue when using k3d and the fabric8 client tries to connect to the cluster an exception is thrown:
 
@@ -111,4 +142,4 @@ the following dependency on the classpath:
   <groupId>org.bouncycastle</groupId>
   <artifactId>bcpkix-jdk15on</artifactId>
 </dependency>
-```
+```