I've got 3 dedicated servers at hetzner.de, and I’m occasionally getting unavailability alerts for these servers, from 2 independent monitoring systems (New Relic and Pingdom) usually for all 3 server at the same time.
Such alerts exceed 5 min, and happens on average once in 2 weeks. Also, in most of the cases, during the reported “downtimes", I can access my servers from home, but can not access from other server in different location - so some requests are coming through and some not, so probably servers are unavailable from only some locations (which happen to be where "pingers" are located).
Hetzner's support repeatedly claims that they didn't have any network issues at the time of those alerts, and that no other customers are complaining about such problems. This doesn’t make sense, especially since the alert are raised simultaneously for all 3 servers, each of which has its own different setup and software installed.
I don’t believe it’s a monitoring bug, because I'm using 2 different monitoring systems. Also, I have servers with similar software configurations in other data centers (not Hetzner), which do not have get any unavailability alerts from these same monitoring systems.
Does anyone have an idea of what might be causing this and/or how I can further investigate the cause to come up with a solution?
