Container performance displayed in grafana is not accurate

[darrylmendillo]

Container metrics recorded and displayed in grafana do not match nomad.

Potential Causes:

• Calculations in grafana query are incorrect
• Scrapping service is not reporting accurate metrics
• Metrics are getting crossed and renamed somewhere between scraping and grafana

Evidence:

Solutions:

• Evaluate query expressions for accuracy
• Step through pipeline to find issues
• Use another docker container metrics tool

[darrylmendillo]

Possibly use nomad metrics instead: https://learn.hashicorp.com/nomad/operating-nomad/prometheus-metrics

This will likely require less compute power than cadvisor:

[darrylmendillo]

Nomad provides prometheus formatted metrics at endpoints. This requires:

• [x] Update nomad agent config and restart all client and server agents
• [x] Evaluate metrics are complete

[darrylmendillo]

Metrics provided by nomad are accurate