Add optionnal simple telemetry so we now how many docs instances are out there

[virgile-dev]

We recently added a table of know instances on the README.
People are encouraged to add theirs via a PR.
Ideally we'd love to be able to know how many instances are out there without having to ask people to do a PR.
The way I see it Docs instances could send a ping and a couple info (Docs version, Number of users, of Docs).
This ping would be optional and could be disabled by setting and env variable to false

[rvveber]

Since we need to get that data from the backend anyways, it would make sense to revive #562 where that data is already collected at the backend and already exposed at a special url.

Then for transmitting this data to you, all you would need to do, is to communicate this special url to your central server.
From there you could query the communicated url and get the data visually displayed in for example a grafana dashboard.

[virgile-dev]

I like it ! @lunika your take ?

[rvveber]

I need to revoke that proposal above.
Prometheus is the wrong tool for global Telemetry.
Securing by IP-Address (which correctly pointed out by @lunika is unreliable in clusters)
A central server wouldn't even be able to get the data, its connection would likely go through a Reverse-Proxy, in which case it would be treated like any outside connection and be blocked.

The initial idea: sending a ping from the application - is better 👍🏼

[virgile-dev]

I created this as a discussion to see what was the community's reaction to this. But it seems it would be better to go back the issue format to start discussing implementation strategies.