- 2 minutes to read

JMX, FAQ, Troubleshooting, Common Questions JMX, FAQ, troubleshooting, common questions 4. How often does the JMX Gateway poll Java applications?

4. How often does the JMX Gateway poll Java applications?

4. How often does the JMX Gateway poll Java applications?

Default: 60 seconds (configurable 30-300 seconds). Gateway polls each JVM's MBeans every interval: MemoryMXBean.getHeapMemoryUsage(), GarbageCollectorMXBeans.getCollectionCount(), ThreadMXBean.getThreadCount(). Polling frequency trade-off:

  • 30 seconds (aggressive): Near real-time monitoring, faster alert response, higher network/CPU overhead, use for critical production JVMs (Boomi Atoms processing real-time transactions)
  • 60 seconds (recommended): Balance between responsiveness and overhead, suitable for most production environments
  • 120-300 seconds (conservative): Lower overhead, acceptable for dev/QA environments, non-critical batch jobs

Nodinite Agent polling: Agent polls Gateway REST API every 5 minutes (default, configurable 1-30 minutes). Agent retrieves aggregated metrics from Gateway (all 100+ JVMs), evaluates thresholds, triggers alerts. Total latency: Worst case 5 minutes + 60 seconds = 6 minutes from heap threshold violation to alert (heap reaches 86% at 2:00:00 PM, Gateway polls at 2:00:30 PM, Agent polls at 2:05:00 PM, alert fires 2:05:05 PM).

For faster alerting: Reduce Agent polling to 1 minute + Gateway polling to 30 seconds = ~90-second latency (suitable for critical production monitoring).

See all FAQs: [Troubleshooting Overview][]