Modify Health Checks

Use the cw-healthctl tool to modify existing health checks. A common initial modification to the default health checks is to add labels so a set of checks can be run on groups of similar nodes.

For example, to add a label to multiple health checks:

cw-healthctl -i check_nvidia,check_gpu_settings update label=gpu

You may want to update the health check fields for a particular check, such as to run a check more often.

1. Review the current interval for the health check:

   [admin@head0]$ cw-healthctl -i check_porterror ls -l
   Health Checks
     check_porterror
       command: check_porterror.py
       fail_percentage: 15
       fail_streak: 2
       interval: 300
       labels: [infiniband]
   

2. Update the interval for the check:

   cw-healthctl -i check_porterror up interval=120
   

You can also modify a health check using a content file. For example:

cw-healthctl -i check_porterror up --content=@checkporterror.yaml

If you are using ARS, use the cw-remedyctl tool to change the remedies associated with the health check.