- 2 minutes to read
How quickly does Nodinite detect when a WSO2 service stops?
Detection speed depends on what stops, and on two settings you control: the polling interval and the Duration. Nodinite monitors WSO2 with the JMX Monitoring Agent, which watches the WSO2 server's JVM, not individual WSO2 services.
When the WSO2 Server or JVM Has a Problem
The JMX Monitoring Agent reads CPU and heap memory from each WSO2 JVM on every poll. The time from a problem to an alert is roughly:
- the time until the next poll (Polling Frequency), plus
- the Duration you set on the CPU or Heap Size thresholds, which is how long the value must stay over the threshold before the state changes (Configuration), plus
- the delivery time of your Alarm Plugins.
A short polling interval and a short Duration give faster alerts, and more alerts for short spikes. When the WSO2 JVM stops or its JMX port can't be reached, the agent can't read values on the next poll, and the Resource shows Unavailable (see Multi-JVM Monitoring).
When a Single Proxy Service or API Stops
The JMX Monitoring Agent doesn't see the state of individual proxy services, APIs, or endpoints. To detect those, add:
- the Web Services Monitoring Agent, which calls the service over HTTP(S) on its own schedule and alerts when it doesn't answer as expected
- Non-Events on events from WSO2 Enterprise Integrator - Logging, which alerts when expected messages stop arriving