Availability & Trust
We know that if you can't log in to your app nothing else works. At web scale in the cloud, failures and outages occur. That's why we have architected Auth0 to be highly resilient to the failure of any of its components with the goal that your users never experience any downtime.
How it works?
At a high level, our availability strategy is rather simple, and yet very effective: we ensure that critical dependencies are redundant, we rapidly detect failures, and our failover is very quick. The Auth0 architecture implements redundant components at all levels. These are:
- DNS
- Datacenter
- Application layer
- Storage
Our High-Availability Strategy
We've done many things to ensure extra availability. An important aspect is how the application is architected: That is, how we manage user sessions, how we partition functionality, how we prioritize the availability of modules, and how we handle transient conditions
Auth0 is designed and built as a scalable, highly available, multi-tenant cloud service.
This highly reliable architecture is combined with solid operational processes and a culture of continuous improvement that constantly refines and improves Auth0 operations.
Consider the following scenarios:
- If one of the VMs is taken down, the load balancer will take it off rotation until it's up again.
- If one of the database VMs is taken down, the secondary VM is elected as primary.
- If there is an outage on the primary region, we will swap the DNS record to start redirecting requests to the second region (in a different location).
Continuous Deployment with No Downtime
We continuously improve Auth0. Whether we have to update a tutorial, optimize the performance of a module, or launch a whole new feature, on a typical day we might deploy three or four times. We designed our deployment infrastructure to do this with no planned downtime.
Auth0 Status
We monitor our infrastructure from over 50 locations all over the world. Our availability is published instantly on status.auth0.com, and if there ever is an incident, we'll let you know immediately through Twitter. You can also check our uptime reports.
