Availability Monitoring

[LinqLover]

(Self-)hosting is connected to a number of concerns:

1. As the hoster, I would like to be informed ASAP when the server stops working (i.e. is no longer reachable).
2. For retrospective purposes, it would be nice to log this information so that I can see a table/visualization of all downtimes with their date/time and duration.

What about performance measuring (mean time to answer)?

There would be two general approaches:

1. Try to detect errors from within the server image: This would provide qualitative insights about the sources of errors (i.e., a transparent ToolSet layer could be added that uses another TelegramBot to notify the hoster about all errors and warnings that arise). On the other hand, if there is any fundamental problem with the image, the VM, the docker container, the raspi, or the network of the hoster (see wiki), this approach won't fail, of course.
2. Set up an external monitoring server that is in the greatest possible distance to the bot server. Hardly to realize due to limited resources, but it would provide an E2E perspective on the available. Thus maybe also relevant in the context of Integration Tests with TelegramClient are broken TelegramBot#4 (I guess we would need a mock telegram client in both cases).

[LinqLover]

Noticeable that this could also help us to identify and to measure issues such as #14!