JanusGraph metrics data having value 0 for most metrics
I have a janusgrpah server with metrics enabled along with jmx.metric enabled, The issue is all the metric starting with
org_janusgraph have a constant 0.
I'm able to see all the metrics with name org_apache_tinkerpop but not the janusgraph ones. Can someone please suggest if I
m missing or how can I enable it.7 Replies
But your JanusGraph Server is processing some traversals?
Yes, we have a throughput of around 9k rpm (for /gremlin api) to the janusgraph cluster.
Which is also shown by the metric
metrics_org_apache_tinkerpop_gremlin_server_GremlinServer_op_eval_OneMinuteRate screenshot attached.
All these requests have queries which fetches data from the backend storage which in this case in BigTable
Ok, can you please also show some metrics that are
0? And how are you accessing the metrics?I'm using
jmx_prometheus_javaagent to expose the metrics which are then scraped using ServiceMonitor from all the janusgraph instances and pushed to prometheus and then using grafana to access these metrics.
Some of the metrics which are 0:
Can see that some metrics are having data also, for example:
Interesting, I just checked in our deployment of JanusGraph and there I couldn't find these metrics at all where you see the value
0 constantly
I'm wondering whether JanusGraph simply didn't report any metrics for those which is why they are still 0 in your case and in our case maybe something filters them out in our infrastructure if they're constantly 0
You've mentioned that you're using BigTable. I'm not sure if that backend even uses the ExecutorService