At GameAnalytics we collect ~2 TB of data per day from 12.000+ mobile games. It is an exciting technical challenge to store, transform, and process this data, so that game developers can get key insights in a few clicks. A while ago, we came close to the limits of our highly effective and extremely specialised backend system. Changing a battle-proven solution is always a bold move, but in our case, the potential gains outweighed the risks.
In the talk, I’ll give an account of the surprises, setbacks, and key learnings from this journey.
– Why we chose Druid
– How we integrate it in the existing cloud infrastructure
– What to expect from Druid, for supporting a flexible and extensible analytical system
– What future improvements are planned
– What were the lessons learned