Fix typos in changelog

`exhausion` → `exhaustion`
`queing` → `queuing`

[JENKINS-32865] Bump instance-identity to 1.5.1

this indirectly upgrades bouncycaslte to 1.54

[FIXED JENKINS-32765] Allow the directory that plugins are exploded into to be changed

Few czech translation strings added.

Updated depCompatWarning to match compatWarning

Added missing translations for Polish language

(#1986 continuation)

Added missing translations for Polish language

Of course the import being used is in a package name that is reserved for artifacts produced by a JSR... and the JSR that the dependency is claiming to come from has not actually published anything... thus whoever produced the artifact has broken the JSR policy... and it is also not safe for us to actually use these annotations as we do not know what a JSR may change on the road to actually providing this... but that is a war for a different PR

[JENKINS-32517] Deprecate Queue#getApproximateItemsQuickly

[JENKINS-32765] Perhaps it makes more sense to make this a property of the PluginManager rather than the classic strategy

- Also expose via servlet context parameter of same name to allow multiple instances of Jenkins in the same servlet container to have different directories
- As we are no longer a constant, change the system property to camelCase

- This really has to be controlled by either a system property or a CLI parameter as we cannot guarantee that any path
  specified in a configuration file in JENKINS_HOME will be valid on another machine where the JENKINS_HOME is mounted
- Obviously Jenkins does not currently support fully starting two Jenkins instances on the same JENKINS_HOME, so this
  fix is to assist in Disaster Recovery (or semi-higher availability) sceanarios where the backup node is being brought
  up. In such cases, the node that failed may not have correctly released all its file handles in the backing NFS share
  and thus could be blocking the recovery node from starting up.
- An additional use case is where the JENKINS_HOME is stored on a remote disk, e.g. a SAN, etc. and the user wants to
  take advantage of the faster local disk to serve the resources from both plugins and Jenkins core. Without this change
  only Jenkins core can be relocated outside of JENKINS_HOME (using the '--webroot=...' command line option).
- Ideally we would introduce this into the extras-executable-war module once the system property has had time to soak
  (although in an ideal world that module would be agnostic of Jenkins and thus it might not be appropriate there)

[FIXED JENKINS-24696] Cannot assume that Jenkins.getInstance() != null when we are running a boot failure script.
(cherry picked from commit 42de07bb)

(cherry picked from commit 09722a98)

(#1986 continuation)

updated changelog for JENKINS-32328

[JENKINS-26466] Decouple AbstractProject from DiskSpaceMonitor

Now that the item list is obtained lock-free the cache is no longer needed.
Besides, the cache could be returning invalid results for requests with different authorization to the one that cached the result (in any case this invalid result is transient).