New Relic Flex integration is not working

I have used newrelic flex to collect gpu utilization using Nvidia smi , I was able to do so for Linux-based ec2 instance, data was showing up after executing nrql query in query builder, all of the sudden data is not coming in newrelic dashboard for gpu utilization. I referred to the post given below — https://discuss.newrelic.com/t/how-to-add-the-nvidia-gpu-metrics-on-the-newrelic-infra-agent-feat-newrelic-flex/143231. please help me get this issue resolved.

1 Like

bump, need help with this

I have used newrelic flex to collect gpu utilization using Nvidia smi , I was able to do so for Linux-based ec2 instance, data was showing up after executing nrql query in query builder, all of the sudden data is not coming in newrelic dashboard for gpu utilization. Not sure what’s wrong. Any help is appreciated. cc: @ssingh7

Hello @ssingh7 ,

Can you test the integration using this documentation:
Troubleshooting Flex

Please reference these 2 flex examples dealing with the GPU monitoring:

smi-gpu-monitoring
gpu-monitoring-dashboard

Hope this helps.

Hello @cconde,
I have done the flex troubleshooting, I am getting the correct response after running sdo ./nri-flex -verbose -pretty -config_path /etc/newrelic-infra/integrations.d/nvidia-smi-gpu-monitoring.yml , along with this nothing potential I can found on debugging, the issue that I am facing currently is the data pushed by nr-flex is not showing up in the nr dashboard when I am running nrql queries . FYI , I tried smi-gpu-monitoring – but the same issue persists i.e. Data is not showing up in dashboard.

@ssingh7 Would you be able to provide the response you got from initiating flex manually from the command line removing any sensitive info? Also could we have you check the infrastructure agent logs to make sure nothing is failing in there for the flex integration?

Please find the output after executing the nri command.

@bvandercar Please find the output after executing the nri command in attachment.

Hello @sbisht1 the output from nri-flex looks fine, therefore I would focus on the debug logs from the infrastructure agent as a next step for troubleshooting. You may also wish to first upgrade your agent to a supported version, since the running versions I can see are version 1.12 and the latest agent is 1.33.2.

Please see the process for generating debug logs here: https://docs.newrelic.com/docs/infrastructure/infrastructure-troubleshooting/troubleshoot-logs/generate-logs-troubleshooting-infrastructure/

In the debug logs, it will show the agent’s attempts to invoke and send results from nri-flex to New Relic.
Any problems negotiating with the endpoint or otherwise may be recorded.