Service monitoring
A very important consideration is what to monitor.
Indeed, it's very important to collect as much data as possible, in terms of metrics and KPIs, as they may reveal interesting trends, and can be very useful if something unpredicted happens. But at the same time, business users are mostly interested in different kinds of metrics and information, such as the number of transactions per second (or per hour, or per day), the amount of money that passes through the platform, the number of concurrent users, and so on.
Hence, there are two different kinds of KPIs to look for, sometimes with a blurred boundary between them:
- Technical information: Things such as the memory used, the number of threads, the number of connections, and so on. These things are useful for sizing and scaling systems and trying to forecast whether our system will perform well or some interventions are needed.
- Business information: Defining what information is business information...
