Operations5 min read
Monitoring that tells you something
Most alerts are noise, and noise trains you to ignore alerts. The four checks worth having, and the rule that keeps the list short.
An alert that fires often and means nothing is worse than no alert, because it teaches you to dismiss the category. The goal is not more monitoring; it is a small number of checks that are always worth reading.
The four worth having
- Is the process running, checked from outside the machine.
- Does it answer - a real connection, not just an open port.
- Is the disk filling, which is the failure with the longest warning and the least excuse.
- Did the backup actually complete, which is the one nobody checks until they need it.
The rule
If an alert fires and there is nothing to do about it, it should not have fired. Either make it actionable or delete it. A list of four checks you read is worth more than forty you filter into a folder.