Keep bumping warning logs for reading process_agent_config.host_ips

[athom]

Output of the info page (if this is a bug)

2020-03-05 05:45:40 UTC | PROCESS | WARN | (pkg/config/viper.go:168 in GetStringSlice) | failed to get configuration value for key "process_agent_config.host_ips": unable to cast <nil> of type <nil> to []string

Describe what happened:
It keep bumping this message for a period of time.

Describe what you expected:
Do not be so noisy

Steps to reproduce the issue:
Run the agent, and observe the logs.

Additional environment details (Operating System, Cloud provider, etc):

AWS ECS using datadog/agent:7-jmx

Additional infomation:
Search the word process_agent_config.host_ips, can find it only shows in the code,
don't see it in any config files. Maybe can clean up, or at least document it somewhere.

https://github.com/search?q=org%3ADataDog+process_agent_config&type=Code

[mgrecar]

I'm seeing this, too. Running on AWS EKS, seems to occur once during Helm deployments. I assume this is just noise due to deployments disrupting things?

[grv231]

Same, I started noticing it as well when I redeployed the agent, similar to what @mgrecar mentioned. Although it's a warning, I was hoping if there is some config change needed on the DD agent deployment. For me, it's happening on EKS, not ECS.

Update - It's coming randomly now, it has been hours since I redeployed the agent on EKS cluster