Well no … there was is/was a bug in CCP’s API Response leading to a DDoS to CCP’s API Server. Those applications who’re taking care of CCP’s cachedUntil timers got screwed up and are permanently requesting the API. This behaviour was leading to too many API requests and screwed some of my numbers leading to various strange rumors on Twitter.
Here’s what the “current” timers in the API look like:
<?xml version='1.0' encoding='UTF-8'?> <eveapi version="2"> <currentTime>2012-07-01 10:31:13</currentTime> <result> [...] <dataTime>2012-07-01 10:31:13</dataTime> </result> <cachedUntil>2012-07-01 10:31:13</cachedUntil> </eveapi>
I guess everybody can see what was the root cause … If every call is outdated as soon as you request the data … well … I hope CCP fixes their stuff soon. I changed my API pollers and fixed the database, so the numbers should be accurate