Ably is designed from the ground up so that customers can legitimately expect 100% uptime. We offer a range of uptime SLAs to our customers ranging from 99.9% to 99.999%, depending on their uptime requirements. We're able to offer this guarantee because:
- Our infrastructure is spread over 15+ isolated datacenters and 175+ edge acceleration points of presence. Each datacenter can operate in isolation if necessary should there be network issues outside of our control.
- Every active channel in the real-time system has a primary & secondary co-ordinator across two datacenters in the same region and at least one primary server in a geographically separate region. This means at any point in time there are at least 3 copies of the system state and data that can be recovered should any server or region go down.
- All account state, authentication keys and tokens, and statistics are replicated across 1 region and 2 availability zones.
- All persisted messages are replicated at least 3 times in at least 2 data centres across at least 2 regions.
- We use DNS latency based routing and health checks to ensure that clients are automatically connected to the locally available data centre. All DNS has a TTL of 60 seconds so a data centre becoming unavailable will within a minute have traffic routed to one of our other 15+ datacenters.
- Our client libraries provide an additional level of DNS redundancy across two registries and domains. If our DNS latency based routing, or even our entire ably.io domain becomes unavailable, the client libraries will find a suitable data centre available on our backup domain ably-realtime.com.