We've wired the Prometheus JMX exporter java-agent into our Kafka processes, we've then set up alerts in Prometheus for several key metrics to monitor cluster health (under replicated partitions/offline partitions, request handler avg idle percent to measure load etc.) - and because the JMX exporter runs within the Kafka process, any sustained failure to scrape also sets off an alert.
Cheers, Liam Clarke On Thu, Dec 19, 2019 at 12:42 AM Miguel Silvestre <msilves...@gmail.com> wrote: > A simple tcp connection to kafka port (9092) should be enough no? > -- > Miguel Silvestre > > > On Tue, Dec 17, 2019 at 8:33 PM M. Manna <manme...@gmail.com> wrote: > > > Jai, > > > > On Tue, 17 Dec 2019 at 17:33, Jai Nanda Kumar <jainand...@idexcel.com> > > wrote: > > > > > Hi, > > > > > > How to perform a health check on a running Kafka server in AWS EC2 > > > server. > > > > > > > SHouldn't this be part of your liveness probe? Or, are you trying to do > > this adhoc (not how folks do it anyway)? > > > > > > > > Thanks and Regards, > > > A. Jai Nanda Kumar, > > > Senior Software Engineer, > > > 9066260799. > > > Confidentiality Notice: This e-mail message, including any attachments, > > is > > > for the sole use of the intended recipient(s) and may contain > > confidential > > > and privileged information. Any unauthorized review, use, disclosure or > > > distribution is prohibited. If you are not the intended recipient, > please > > > contact the sender by reply e-mail and destroy all copies of the > original > > > message. > > > > > >
