At the same time as releasing our new website and domain todo.vu, we also snuck in a significant update to our server infrastructure.
We have now upgraded our load balancing systems to be more redundant and provide much faster and more sensitive fail-over responses. And we have increased the capacity of our server farm by another 100%.
This is one of those changes that should hopefully go almost unnoticed, but in reality it is a very significant thing for us. While our systems were always load balanced and redundant, there were still edge cases where server outages could affect some users for short periods of time. We have now removed several points of potential failure and improved the monitoring and speed of failover.
We have also improved server logging to make it much easier to identify causes of failure and instances where specific user activity is causing excessive loads.