I’m using a GH instance as a routing server with an MMAP_STORE_SYNC storage. Today I did some stress tests and observed that apparently, some long distance route requests (~100km as the crow flies) seemed to have caused GH not responding well to parallel short distance requests. e.g. the log trace showed that some requests seem to have run for about 30 minutes.
Now I have a rather low-end V-Server so I’m not surprised that it doesn’t scale well, but I need to think on how to best “protect” the server against such requests. One option would be to do a quick haversine distance calculation between route points and introduce a threshold value beyond which GH isn’t called anymore (wouldn’t be a problem because the engine is intended smaller distance foot/bike routing anyway). But this is a sort of a hack.
So my question is if there is some way to define a timeout in the GH server so that it terminates requests after a specified time?