We need to set up some basic monitoring on the staging and production registries. Here are some ideas for things to monitor:

• Disk space on /var/lib/registry on each registry node.
• Memory usage on each registry node.
• The /v2/ API call on the load balancer, i.e. registry.fedoraproject.org and registry.stg.fedoraproject.org. An HTTP GET /v2/ should return an empty JSON object. For example,

$ http https://registry.fedoraproject.org/v2/
HTTP/1.1 200 OK
Accept-Ranges: bytes
Age: 0
AppServer: proxy04.fedoraproject.org
AppTime: D=145257
Connection: Keep-Alive
Content-Length: 2
Content-Type: application/json; charset=utf-8
Date: Mon, 20 Mar 2017 18:18:29 GMT
Docker-Distribution-Api-Version: registry/2.0
Keep-Alive: timeout=15, max=500
Server: Apache/2.4.6 (Red Hat Enterprise Linux)
Strict-Transport-Security: max-age=15768000; includeSubDomains; preload
Vary: Accept
Via: 1.1 varnish-v4
X-Varnish: 6380586

{}

• The /v2 API call on each registry node, i.e., docker-registry01.stg.phx2.fedoraproject.org, docker-registry02.stg.phx2.fedoraproject.org, the production registries, etc.

This effort is tracked by this Taiga card:

https://taiga.fedorainfracloud.org/project/acarter-fedora-docker-atomic-tooling/us/757