fix #1349, JobInterruptMonitorPlugin memory leak and concurrency issues

[e210]

This PR fixes three bugs in JobInterruptMonitorPlugin as reported in #1349.

Fixes issue #1349

Changes

• Set setRemoveOnCancelPolicy(true) on ScheduledThreadPoolExecutor so cancelled tasks are immediately removed from the work queue, preventing memory leaks
• Replaced the single ScheduledFuture field with a ConcurrentHashMap<String, ScheduledFuture> keyed by fireInstanceId, so concurrent jobs no longer overwrite each
  other's monitor
• Changed getJobDetail().getJobDataMap() to getMergedJobDataMap() so that AutoInterruptable and MaxRunTime can be configured on the trigger as well as the job
• Fixed debug log to report the actual delay used instead of always logging the default value

---

Checklist

• tested locally
• updated the docs
• added appropriate test
• signed-off on the DCO referenced in the CONTRIBUTING link below via git commit -s on my commits, and submit this code under terms of the Apache 2.0 license and
  assign copyright to the Quartz project owners

---

In submitting this contribution, I agree to the terms of contributing as referred to here:
https://github.com/quartz-scheduler/contributing/blob/main/CONTRIBUTING.md

[melloware]

Thank you for the PR

[e210]

Rebased onto latest main (9294eac4) and force-pushed — the diff is unchanged from the version you approved, only the base commit moved.

The earlier red build wasn't caused by this PR: every build in that window failed (#1455, #1466–#1469 as well) and the pipeline recovered on its own by Feb 23. All checks are green now.

@melloware this one has been sitting since February — anything else needed from my side?

[melloware]

No it looks good. I don't have commit permissions only review and tagging permissions. Only @jhouserizer can merge PRs