Game Server status monitor page

Posted by

After various outages and problems we've had in the last week or so (which seems to have improved after moving to a faster server, yay!) I decided to add a small status page to the website where you can go and check whether the server is down for everybody, or if maybe it's just something local to you. So now if you visit, you'll see something like this:

Here you can see we have a brief outage last night when I accidentally detactched the disk which had the database on it (whoops!) but the good thing about this new status page is that it also emails me when it can't contact the server (or when the server returns an error).

The way it works is, every 5 minutes, I do a request to the game. The request does a bit of database reads, a bit of CPU and a bit of database writes. This will hopefully test everything that normal requests will do, so that it gives a good indication of what you can expect from the actual game server. Then it displays the last 24 hours in this nice graph format. At some point, I might start doing more historical data as well, but for now at least, the last 24 hours is it.

One interesting point here is the spikeyness of the response times. It seems like there's a bit of a spike in activity around midnight (Sydney time: all the times are local, and I'm in Sydney), which is OK, but I'm not sure what's causing the spikes at the other times, which all seem to happen fairly regularly... something to investigate!

blog comments powered by Disqus