Several hundred people, myself included, will have noticed that there was nothing to see here for about thirteen hours yesterday and through until the early hours of this morning. The site was available on and off but, officially, completely off-line. Why? Well, from my point of view because of an extraordinary amount of traffic which overloaded the CPU.
I say from my point of view because those in charge of the server – and I don’t mean me as webmaster – seem to feel that there was nothing out of the ordinary at all. That, at least, is the response I got by mail when we telephoned and asked them to check what was happening. And that response came about ten hours after we called in. To say that the worker on the telephone was totally disinterested is only part of the problem, he asked us how we knew that the site was suffering from extreme traffic; information which is available on the configuration page of the host themselves, and easily available to him. The graph showed an increase of traffic from a normal of about 50MB to a constant high of about 4GB. Not the sort of thing that you can easily overlook.
Image Source: Javier Aroche – Creative Commons
What he didn’t pass on to the techs – who should be in the know, one would have thought – was that I was getting a constant 500 error – internal server error – and a report that the CPU was overloaded, and a report that Port 80 had closed down, making the server unavailable for the outside world.
The answer I received, based on the lack of information this worker passed on, was that the host had recently changed the accounting system for traffic – all internal traffic is now included in the end sum. This change took place on 4 July, which was not yesterday. The problems began at nine thirty yesterday. And even if the internal traffic is included, a jump to 4GB?
I think not.
We are now back, as you can see. Hopefully it will remain that way.
- Viktoria Michaelis.