I'll be watching you - Monitoring is everything

Imagine

Imagine a perfect world, with your perfect system, when everything goes fine. And fine. More fine […] And then suddenly you have a call from your boss/client/colleague:

Why our button XYZ is so slow? What’s going on?

Then you check:

1. IIS/apache is fine
2. Database is fine
3. JavaScript looks like normal
4. External functionality is good

So where is the problem? You check XYZ button and it is fine.

So what just happened?

The sad answer: now it is nothing. The problem was 2-3 hours ago and it is gone now. You just spent some of your time completely useless. The question which suddenly comes to your mind is: what had happened and what changed? Moreover one more should appear: when it will happen again?

Monitoring

There are a lot of tools, you should use to measure your application:

1. Performance counters or JMX - this is the base
2. Logs - unfortunate most developers (including myself) create poor quality logs. But quick access to logs is very useful. Moreover try ELK or Splunk enable fast search over history
3. Live dashboards - I know that Zabbix or Nagios are old and ugly, but they are damn useful.
4. Add more measures to you code with tools like MiniProfiler/Glimpse on .NET or Metric in Java. This can show use easy bottlenecks in our applications. Moreover we can see it live with tools from point 2 and 3.

We are IT, we are lazy, we need tools which will do work for us. Just find what is useful for you and check what your admins use :)

To sum up - remember Sting song :

Every move you make Every vow you break Every smile you fake every claim you stake I’ll be watching you