In #4499 we added a liveness probe and had the agent properly support reloading the HTTP monitoring server settings so that Fleet managed agents that defaulted to having the server off can turn it on.

This however has caused several unexpected problems, because there are agents (including the agent used to run Fleet Server in ESS) that have enabled HTTP monitoring in the initial agent configuration expecting it to stay enabled. When we enabled reload, what happened is the agent began respecting the default fleet policy which has HTTP monitoring disabled by default, thus turning off the HTTP server with no simple way in Fleet to turn it back on.

As a short term solution while we figure out how to make the default settings work correctly (I believe for ESS we need a way for the default fleet policy used to specify HTTP monitoring server settings), we need to keep the new liveness probe but disable reloading the HTTP monitoring server settings again.