Infinispan Monitoring
Internal
Relevance
- JDG 7
Overview
This article describes choices of metrics for Infinispan production monitoring.
Metric Collection
HotRod Endpoint
Bytes read count and bytes written count, which can be used to infer the read and write rate, in bytes/second (JDG 6 JMX totalBytesRead, totalBytesWritten, JDG 7 JMX bytesRead, bytesWritten. It is preferable to expose these metrics as a rate per second. Some monitoring solutions allow the rate to be calculated automatically (see Data Dog counters).
The number of worker threads (JDG 6 JMX numberWorkerThreads: JDG 7 JMX workerThreads).
The Cache Container
The metric at cache container level represent an aggregation of the metrics for the underlying cache and in general, are a good indicator of the activity on that specific cache node - normally, there's just one cache container.
Interesting metrics are:
- the number of hits, misses and stores at the cache container level (JDG 7 JMX hits, misses, stores). This is an aggregation of the corresponding values for the underlying caches. It is preferable to expose these metrics as a rate per second. Some monitoring solutions allow the rate to be calculated automatically (see Data Dog counters).
- the total number of entries at cache container level. This is an aggregation of the corresponding values for the underlying caches. For more details on the cache number of entries see Infinispan Cache Number of Entries.
- the average read, write and remove time for the caches managed by this container.
Individual Caches
TODO.