High Availability

So your're chasing after those 9's, but how do you know? If your systems have never failed and until now have run consistenly, what's to stop you from your current streak? But you're on a cloud provider, right? Or have at least run in multiple availability zones/datacenters? Are you ok in the event of a DNS failure? Is BGP routing correctly?

These are all questions we ask when looking for those always on services.

Every cluster in the Latency network is created with a specific region assigned. This assignment is a configured in all our software which is replicated in each datacenter. Beyond replication we also have static IP assignment for failover scenarios. Our service is actively monitoring BGP routes....

Chaos Availability Networking

In order to put our resilincy to the test, once a week, latency will completely delete a cluster piece by piece until it no longer exists and then rebuild it. The cluster is chosen at random and is indicated on the root of our API.

Note: This project is in early alpha. Clusters may be shut down longer than normal in unexpected circumstances.