New in this release
From this release UCS nodes are self-monitoring, and can send alarms, as follows:
UCS will send alarms if:
- UCS process CPU is more than 90% for one minute (LMS 20000)
- JVM heap usage is more than 90% for one minute (LMS 20002)
- GC freeze is more than 2 seconds (LMS 20004)
- Mean time spent in ESP queue is more than 500 milliseconds (LMS 30101)
- Mean time spent processing ESP requests more than 1 second (LMS 30103)
- UCS thread pool size is 200 (LMS 30105)
- Mean response time from UCS 8 is more than 5 seconds (LMS 30107)
- Cassandra mean response time is more than 2 seconds (LMS 30109)
- Cassandra monitoring should be set up according to Instaclustr recommendations.
- Elasticsearch monitoring should be set up according to Elastic recommendations.
This page was last modified on December 28, 2018, at 16:14.