[doc] Add known issue to all versions affected by GC behaviour change (#102025)

docs/reference/release-notes/8.10.0.asciidoc

@@ -35,6 +35,8 @@ delete all the snapshots in the repository taken with version 8.10.0 or later
 using a cluster running version 8.10.4.
 // end::repositorydata-format-change[]
 
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 [[breaking-8.10.0]]
 [float]
 === Breaking changes

docs/reference/release-notes/8.10.1.asciidoc

@@ -9,6 +9,8 @@ Also see <<breaking-changes-8.10,Breaking changes in 8.10>>.
 
 include::8.10.0.asciidoc[tag=repositorydata-format-change]
 
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 [[bug-8.10.1]]
 [float]
 === Bug fixes

docs/reference/release-notes/8.10.2.asciidoc

@@ -7,4 +7,6 @@
 
 include::8.10.0.asciidoc[tag=repositorydata-format-change]
 
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 Also see <<breaking-changes-8.10,Breaking changes in 8.10>>.

docs/reference/release-notes/8.10.3.asciidoc

@@ -7,6 +7,19 @@
 
 include::8.10.0.asciidoc[tag=repositorydata-format-change]
 
+// tag::no-preventive-gc-issue[]
+* High Memory Pressure due to a GC change in JDK 21
++
+This version of Elasticsearch is bundled with JDK 21. In JDK 21
+https://bugs.openjdk.org/browse/JDK-8297639[Preventive GC has been removed].
+This may lead to increased memory pressure and an increased number of CircuitBreakerExceptions when retrieving large
+documents under some particular load. (issue: {es-issue}99592[#99592])
++
+If you needed to explicitly <<known-issues-8.10.0,enable Preventive GC for your workload in a previous Elasticsearch
+version>>, we recommend you avoid to upgrade to this version, as the settings to enable Preventive GC have been removed
+from JDK 21.
+// end::no-preventive-gc-issue[]
+
 Also see <<breaking-changes-8.10,Breaking changes in 8.10>>.
 
 [[bug-8.10.3]]

docs/reference/release-notes/8.10.4.asciidoc

@@ -25,6 +25,8 @@ first. If you cannot repair the repository in this way, first delete all the
 snapshots in the repository taken with version 8.10.0 or later using a cluster
 running version 8.10.4.
 
+include::8.10.3.asciidoc[tag=no-preventive-gc-issue]
+
 Also see <<breaking-changes-8.10,Breaking changes in 8.10>>.
 
 [[bug-8.10.4]]

docs/reference/release-notes/8.11.0.asciidoc

@@ -10,6 +10,11 @@ Also see <<breaking-changes-8.11,Breaking changes in 8.11>>.
 Infra/Core::
 * Remove `transport_versions` from cluster state API {es-pull}99223[#99223]
 
+[[known-issues-8.11.0]]
+[float]
+=== Known issues
+include::8.10.3.asciidoc[tag=no-preventive-gc-issue]
+
 [[bug-8.11.0]]
 [float]
 === Bug fixes

docs/reference/release-notes/8.7.1.asciidoc

@@ -18,6 +18,23 @@ This issue is fixed in 8.8.0.
 
 include::8.6.0.asciidoc[tag=reconciliation-imbalance-known-issue]
 
+// tag::no-preventive-gc-issue[]
+* High Memory Pressure due to a GC JVM setting change
++
+This version of Elasticsearch is bundled with JDK 20. In JDK 20
+https://bugs.openjdk.org/browse/JDK-8293861[Preventive GC is disabled by default].
+This may lead to increased memory pressure and an increased number of CircuitBreakerExceptions when retrieving large
+documents under some load patterns. (issue: {es-issue}99592[#99592])
++
+If this change affects your use of Elasticsearch, consider re-enabling the previous behaviour
+by adding the JVM arguments `-XX:+UnlockDiagnosticVMOptions -XX:+G1UsePreventiveGC` (reference:
+https://www.oracle.com/java/technologies/javase/20-relnote-issues.html#JDK-8293861[JDK 20 release notes]). It is
+important to note that this workaround is temporary and works only with JDK 20, which is bundled with Elasticsearch up
+to version 8.10.2 inclusive. Successive versions are bundling JDK 21+, where this setting
+https://bugs.openjdk.org/browse/JDK-8297639[has been removed]. Specifying those JVM arguments will prevent the
+JVM (and therefore Elasticsearch Nodes) from starting.
+// end::no-preventive-gc-issue[]
+
 [[bug-8.7.1]]
 [float]
 === Bug fixes

docs/reference/release-notes/8.8.2.asciidoc

@@ -3,6 +3,11 @@
 
 Also see <<breaking-changes-8.8,Breaking changes in 8.8>>.
 
+[[known-issues-8.8.2]]
+[float]
+=== Known issues
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 [[bug-8.8.2]]
 [float]
 === Bug fixes

docs/reference/release-notes/8.9.0.asciidoc

@@ -12,6 +12,8 @@ task is longer than the model's max_sequence_length and truncate is set to none
 then inference fails with the message `question answering result has
 invalid dimension`. (issue: {es-issue}97917[#97917])
 
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 [[breaking-8.9.0]]
 [float]
 === Breaking changes

docs/reference/release-notes/8.9.1.asciidoc

@@ -3,6 +3,11 @@
 
 Also see <<breaking-changes-8.9,Breaking changes in 8.9>>.
 
+[[known-issues-8.9.1]]
+[float]
+=== Known issues
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 [[bug-8.9.1]]
 [float]
 === Bug fixes

docs/reference/release-notes/8.9.2.asciidoc

@@ -3,6 +3,11 @@
 
 Also see <<breaking-changes-8.9,Breaking changes in 8.9>>.
 
+[[known-issues-8.9.2]]
+[float]
+=== Known issues
+include::8.7.1.asciidoc[tag=no-preventive-gc-issue]
+
 [float]
 [[security-updates-8.9.2]]
 === Security updates