In the Software Engineering world we have seen the value of automating the monitoring of the systems.
For example we can monitor when a website is not responding to a https request and then alert somebody to go and do something about it, before a user/customer finds the website is down.We can apply the same thinking in the data space.
We can automate the monitoring of when a load happens or more importantly when it doesnt.
This automated monitoring is even more important given the fact that the Modern Data Stack is the cobbling together of a stack of seperate products/technologies.
Automated monitoring can provide insight into which of the many jenga blocks caused a failure so we know where to start the remediation.