Cassandra Monitoring:
- Disk / IO latencies
- CPU utilization
- Memory
- JVM metrics
And all needed metrics to monitor Cassandra from nodetool outputs?
Thanks,
-Varun
Bringing together the Apache Cassandra experts from the community and DataStax.
Want to learn? Have a question? Want to share your expertise? You are in the right place!
Not sure where to begin? Getting Started
What are the best tools out there to monitor Cassandra servers?
The Best tool for DSE is OpsCenter.
But there are many thirparty tools which you can configure like :-
-Datadog
-grafana
-Dynatrace
-APM tools
-Influx
-Telegraf
-etc, etc
@vkayanala_42513 In DSE 6.7, we introduced the new DSE Metrics Collector which collects metrics data that you can push to monitoring systems like Prometheus and Graphite. For more info, see Improved Performance Diagnostics With DataStax Metrics Collector. Cheers!
Thanks for responding to my question.
@Erick Ramirez I'm following this document: https://docs.datastax.com/en/dse/6.7/dse-admin/datastax_enterprise/tools/metricsCollector/mcExportMetricsManually.html to get Cassandra metrics into Prometheus and then display onto Grafana dashboards.
cont..1
I'm facing some challenges in this process.. both Prometheus and Grafana services are up and running, but It can't able to reach dse servers endpoints you can see that in below screenshot.
9103 port opened on these servers what else I need to verify here to get status UP? And also data is not populated on Grafana too?
What I'm missing here ;) Any help would be appreciated!
-Cheers!
@vkayanala_42513 can you verify that the DSE nodes are listening on port 9103?
@Erick Ramirez 9103 port is opened on dse servers. But 9103 used for what service? should we start any service to make it listen on?
Output from one of the dse server:
[root@cassandra-perf-dev-005-db ~]# netstat -lnt | grep 9103
[root@cassandra-perf-dev-005-db ~]#
@vkayanala_42513 it looks like you missed step 2 of the document and didn't disable/re-enable the DSE Metrics Collector. If you did, the DSE nodes would be listening on port 9103. For example:
$ sudo lsof -i -n -P | grep LISTEN | grep 9103 ld-2.23.s 7601 cassandra 5u IPv6 31579 0t0 TCP *:9103 (LISTEN)
@Erick Ramirez Yeah you are right, in step two, prometheus.conf file was placed in wrong directory. Now target end points status is UP. Prometheus is good.
However I'm notable to getting those metrics into Gafana. I noticed this on Grafana dashboards "Datasource named prometheus was not found".
You might like to look for the following open source tools, which can expose the jmx data. Later you can use anything on the top. Put the data in influxdb + plot on Grafana would be easy :).
https://github.com/jmxtrans/jmxtrans
and Telegraf with Jolokia2
https://github.com/influxdata/telegraf/tree/master/plugins/inputs/jolokia2
I don't know much, and still it's a learning curve. However, the following few data might help, and if some else can share there learning, it will be helpful.
EDIT: Moved large amount of text to Gist since it made the post hard to read.
In our env, we use grafana+influxdb+telegraf. it is pretty handy. It is open source free. We can almost get all the key metrics. You can take a look if you are interested here.monitor Cassandra with open source tool
3 People are following this question.
