Before going into the mess of authentication and data that needs to be read from and written to potentially complex data-stores, it would be nice to be able to just ping the server and know that its okay.
Thinking ahead, this would probably also be something nice for an admin that got a page at 3 in the morning that something is wrong.
Pinging the server is really simple, we’ll just send a ping request to the server, and it should send back a reply of whether or not its healthy, and a string with further information.
On the server we’ll do this as a module that has an internal value that is updated atomically and protected by a mutex, and then have a set of methods that can set this value to true or false based on if the server is healthy. The idea here is that its really hard to tell if the server is good, but its easy to tell if something bad has happened that’s unrecoverable. Once an unrecoverable error has occured, the system should be able to expose that not only in the logs, but in the ping response.
In the future this will likely be fleshed out to a set of system specific health checks, but for now a single global status will be more than enough to keep things going.>> Home