I love strace, DTrace, Procmon, and all these tools. Nothing gives me more pleasure when I'm called to solve a mistery issue and end up spending time checking s

I love strace, DTrace, Procmon, and all these tools. Nothing gives me more pleasure when I'm called to solve a mistery issue and end up spending time checking syscalls, timing, etc. That being said, that's the worst debugging experience your company can have because it outlines a bigger problem: your monitoring sucks. In the scenario presented by this article, better monitoring of the database layer could have allowed the sysadmin to connect the dots much faster and find the root cause. So the DB might be able but spilling the wrong beans? You may need some functional monitoring checking what it's outputting from time to time. I absolutely love low-level debugging tools but using them on a daily basis doesn't scale. I do use them almost every single day, because other teams failed to properly monitor and understand their systems. That doesn't mean it's the most efficient way to go about this.