Telegraf: Add support for kubernetes jobs

[carlosv-ukg]

kubernetes Jobs

Opening a feature request kicks off a discussion.

Proposal:

Add functionality to generate metrics derived from the state of the job resource to the kube_inventory input plugin. This could be a new metrics like:

• kubernetes_job
  
  
  
  • tags:
    
    
    
    
    
    • job_name
    
    
    
    • namespace
    
    
    
    • selector (*varies)
    
    
    
  
  
  • fields
    
    
    
    
    
    • started
    
    
    
    • completed
    
    
    
    • failed
    
    
    
    • succeeded
    
    
    
  
  

Current behavior:

Job metrics are not currently ingested.

Desired behavior:

Ingest job metrics.

Use case

By bringing in job metrics into influx, a user can:

1. Set up alerting when jobs start failing which allows users know about issues as soon as possible without having to monitor jobs.
2. See the health of jobs over time and see if there is an issue with applications or infrastructure.

[ssoroka]

I really like this. I'd love to see it.