Can a service even have a 99.999% uptime guarantee? It's easy to promise, but actually delivering on that is another challenge entirely. Testing, CI/CD, reviews--strategies to achieve three nines are no longer sufficient.

Running critical components requires a completely different mindset when the required uptime is five nines, and there are many reasons to require high uptimes--regulatory needs, global application operations, or in some situations your service could offer life-saving responsibilities. In these circumstances, it becomes critical to get this right, and never let your service go down.

In this talk, I'll dive into if five nines is even possible, a full review of the challenges we encountered building our mission critical service Authress, and iterate through the key trade-offs in elevating the reliability of our services.

Key Takeaways:
* The core components of a highly reliable solution
* Lessons learned in the process
* Understanding service reliability in the business context
* Architecture strategies to increase the reliability