Key performance indicators and a focus on the mean time to recovery are all well and good, but they are also largely academic unless there is traffic running through your cloud-native system. International companies will likely have traffic around the clock, but many systems do not have an international user base. Furthermore, even though the system as a whole may be receiving continuous traffic, individual regions will receive different levels of traffic throughout the day.
We are continuously deploying changes into production and we need to proactively assert the health of the system immediately after each deployment to be confident in the success of the deployment. However, without traffic in the system, there is no information available to assert the health of the system. This is particularly true when we are performing a canary deployment in an...