https://datahubproject.io logo
Join Slack
Powered by
Is it possible to enable metrics for datahub-front...
# all-things-deployment
f
Is it possible to enable metrics for datahub-frontend via helm chart, i can see an option to enable jmx exporter.
b
@orange-night-91387 or @incalculable-ocean-74010 can help with questions on our helm charts!
i
JMX exporter gives Java based metrics like GC times, heap sizes and so on.
Simply set enabled to 
true
Which you can get by querying 
/metrics
f
What kind of metrics datahub-frontend is exposing? And how to enable it?
And where is env.JMXPORT is being used?
i
cc @bulky-electrician-72362
b
hey, JMX is not installed on the frontend pod at the moment. I'm working on the install and releasing it probably this week.
also the jmx exporters in general doesn't work properly in the helm chart (fix is in PR phase). if you want metrics about the gms pods for example:
Copy code
global:
      monitoring:
        enablePrometheus: true
set this in the helm chart and you can set prometheus to scrape it by:
Copy code
values:
    datahub-gms:
      enabled: true
      podAnnotations:
        <http://prometheus.io/scrape|prometheus.io/scrape>: "true"
        <http://prometheus.io/path|prometheus.io/path>: "/metrics"
        <http://prometheus.io/port|prometheus.io/port>: "4318"
f
Also looks like jmx side car container option that was provided in helm chart doesn't have any use right?
b
yes exactly, there is JMX JVM agent in the container that's what needs to be enabled/scraped. \
f
We have jmx agent in gms, mae, mce only right?
b
Yes, forntend is coming soon (it's the task I'm working on)
f
Thanks
Is it working for mce-consumer?, I don't see any metrics for it.
b
mce-consumer itself is broken on the main branch at the moment. which version are you on? Is the pod runinng for you?
f
0.8.41
b
if
Copy code
global:
      monitoring:
        enablePrometheus: true
is set and the pods are running then you see the metrics on 
4318/metrics
f
I can see for gms, mae but not mce
b
f
Yes, But I'm not getting metrics at 4318/metrics while I’m able to get it for gms, mae
b
are there any errors warrning from mae that can hint us what's happening?
can you see the PROMETHEUS_ENABLE env vars in the env vars in that pod? (you can list env vars with the 
env
command)
f
I'm using extaEnvs to add ENABLE_PROMETHEUS in mce and it's working now. THANKS!
b
anyways, glad it works for you
f
We are using datahub-0.2.87 which doesn't have these lines. It is fixed in the latest versions
Open in Slack
PreviousNext
Powered by