High availability

General Site Updates
1

giphy

We finally have High availability

Specifically the forum has a fault tolerance of 12 second on the web nodes.
This means I can make changes to the forum without taking it down.

When an outage happens on one of the web nodes and you refresh within that 1 second window, a connection will fail. But refresh again, and you’ll be directed to the up node.

I will be monitoring the server over the next few days to make sure things remain stable but I have no reason to believe anything bad will happen.

24 Likes
2

Just to update a bit on this.
I occasionally get emails that the health check heartbeat thinks nodes go down because 1 second is a bit too fast for full round packet responses. Reduced to 2 second fault tolerance to avoid flipping the load back and forth between nodes.