refactor: Enforce 15-char limit and simplify labels for thread naming (#6212)

This change continues the thread naming work from #5691 and #5758, which enables more useful lock contention profiling by ensuring threads/jobs have short, stable, human-readable names (rather than being truncated/failing due to OS limits). This changes diagnostic naming only (thread names and job/load-event labels), not behavior.

Specific modifications are:
* Shortens all thread/job names used with `beast::setCurrentThreadName`, so the effective Linux thread name stays within the 15-character limit.
* Removes per-ledger sequence numbers from job/thread names to avoid long labels. This improves aggregation in lock contention profiling for short-lived job executions.

src/xrpld/app/main/NodeStoreScheduler.cpp

@@ -12,9 +12,8 @@ NodeStoreScheduler::scheduleTask(NodeStore::Task& task)
     if (jobQueue_.isStopped())
         return;
 
-    if (!jobQueue_.addJob(jtWRITE, "NodeObject::store", [&task]() {
-            task.performScheduledTask();
-        }))
+    if (!jobQueue_.addJob(
+            jtWRITE, "NObjStore", [&task]() { task.performScheduledTask(); }))
     {
         // Job not added, presumably because we're shutting down.
         // Recover by executing the task synchronously.