We all know, probably since forever, that we always need to design our solution to be highly available, resilient, fault-tolerant, etc. But, does it mean that if the solution is resilient on paper, it will also be resilient in real life? We may easily make a mistake during the design phase, development phase, operation phase, etc. And when your solution is down, and the business people are breathing down your neck, it is easy to make yet another mistake and make everything even worse than before.
Implementing a monitoring solution is good, but it will just help you to understand faster that you are screwed. Having a disaster recovery plan is awesome, but how can you be 100% sure that everything will go according to plan? The only way of doing that is to try it upfront in a controlled environment.
During my session I will be talking about planning and emulating your infrastructure failures in a controlled environment, which will allow your team to run disaster recovery drills and be prepared for any kind of nasty surprises.
