What is monitoring?
Monitoring shows whether servers and services are up, and how busy they are.
What it is
Monitoring answers questions once the app is running, not while you write the code. Is the service reachable? How full are CPU, memory, and disk? How long do requests take? Is the error rate rising?
It answers the question while the system is running: is the service up, and how full is the machine it runs on.
What it is made of
There are two views. Availability checks from outside whether an address still answers. Load looks at the machine and the process itself. Together they tell you whether the site is only slow or already gone. A check that only asks “does the port answer” does not notice when the app is stuck and still keeps the connection open.
What goes wrong
Thresholds and a message belong to it. A graph nobody looks at is an archive. A message for every small hiccup is noise. You decide what a real problem is: disk over a value, service unreachable for more than a few minutes, error rate above the usual.
What you otherwise look after yourself
Without monitoring you notice outages when someone writes. At night or on a weekend that is often too late, and you do not know since when it has been broken.
How it shows up on MainPath
On MainPath monitoring is part of operations, next to logs and backups. You see whether servers and services are still there, without building a separate check chain for every project.