cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Oleksandr Shulgin <oleksandr.shul...@zalando.de>
Subject Decommissioned nodes and FailureDetector
Date Fri, 19 Jan 2018 08:42:41 GMT
Hello,

Is there a better way to monitor for Cassandra nodes going Down than
querying via JMX for a condition like FailureDetector.DownEndpointCount > 0?

The problem for us is when any node is decommissioned, it affects the
DownEndpointCount for another ~3 days (the famous 72 hours of gossip).

Is there a similar metric to be observed which doesn't include nodes which
are expected to be down?

Regards,
-- 
Oleksandr "Alex" Shulgin | Database Engineer | Zalando SE | Tel: +49 176
127-59-707

Mime
View raw message