Implement a monitoring job for workers

There is a nice, simple monitoring service: https://uptime.kuma.pet.

One can set custom ways to monitor a service: make http requests (pull based) or wait for a special HTTP request (push-based).

One can also set custom ways to get notifications: email, telegram, etc.

Currently, we can monitor the website of GGTX just by making a GET request on https://dev.sub.gargantext.org

However, for workers, one should go with a push-based solution.

I propose to add a http push job to Gargantext. Such a job would have a small timeout (as it just makes a http request, it should be fast) and would have strategies set up so that it is always restarted (with an appropriate delay, say, 60 seconds) -- all of this is possible with the current haskell-bee already. This job would make a HTTP push request to some kuma endpoint, this can be configured in TOML.

This way we know that workers keep processing and there isn't a large delay. This would be helpful to debug issues like e.g. purescript-gargantext#756 (comment 15184) where I'm not sure if it's a bug in code or just workers not running.