Pulp

When using the /status/ API to evaluate the health of a Pulp system, webserver availability and health information is not included. In a standalone installation this is not needed because the response code of 200 OK is intended to indicate the webserver is OK. In a clustered environment with multiple webservers, a /status/ request may be handled by a webserver who is online and functioning, while other webservers may be offline.

To workaround this limitation, it is recommended that all webservers in a cluster have the /status/ API call monitored directly (not through the load balancer). All webservers that are expected to be online should be checked that they return a '200 OK'.

This is likely not something Pulp can resolve currently because the load balancer is the component that knows which webservers to expect to be online or not. I propose this bug remain open to track this issue as a known limitation.