Enable TaskMonitor warnings of slow tasks by default

HBase has a TaskMonitor that can warn on any tasks that take too long.
https://github.com/apache/hbase/blob/47996d6c2128815e45bb8bdb6e3a470bfddd6106/hbase-server/src/main/java/org/apache/hadoop/hbase/monitoring/TaskMonitor.java#L49

By default it does not warn as the setting hbase.taskmonitor.rpc.warn.time defaults to 0 which disables the check. But it's definitely useful to have a warning like this.

We saw tasks stuck for a week at a customer and this went unnoticed. As I'm not 100% sure on which tasks (i.e. also Procedures?) are monitored I suggest a rather high threshold.
Maybe 1 hour? (Even though most tasks should finish within seconds)

This is a warning that we then - later - should extract from the logs to convert into an alert and/or at least have a dashboard (in Grafana or Opensearch Dashboard) to show these.
The log message will start with: "Task may be stuck"

Also note that older versions have a bug that can wrong warnings: https://issues.apache.org/jira/browse/HBASE-22935

Beginne mit der verknüpften hbase-server/src/main/java/org/apache/hadoop/hbase/monitoring/TaskMonitor.java und untersuche, wie hbase.taskmonitor.rpc.warn.time die Prüfung deaktiviert. Bestätige, welche Tasks überwacht werden, und lege den Standardgrenzwert fest, etwa eine Stunde. Als abgeschlossen gilt die Aufgabe, wenn Warnungen zu langsamen Tasks standardmäßig aktiviert sind, ohne das ältere Verhalten mit Fehlwarnungen, das in HBASE-22935 beschrieben wird.