Anyone who looks after several applications as a developer or small agency knows the situation: each project runs on a different server, is built differently and has a different contact person. When one fails, you often hear about it from the customer.
Five projects, five requirements
- Shipping software that sends labels to printers through an interface. What counts is that the interface answers, at night too.
- An online shop for trade products. Orders arrive at any time; a shop that fails silently costs revenue.
- A small web application with few users, where every follow-up question costs time.
- An employee portal for leave, documents and surveys. Staff expect it to work in the morning.
- Occupational-safety software with which specialists document on site at their clients. An outage affects several people at once.
One monitor per project to start
To start, one HTTP monitor per project is enough, pointing at an address that only returns status 200 when the application works. Special cases come on top: a port monitor for a service without a web page, a DNS monitor for a domain whose renewal nobody should forget. The Free plan allows ten monitors, plenty for five projects.
The real gain
Instead of keeping in mind for every project whether it was running last time, there is one dashboard with all monitors at a glance. When everything is green, there is time for actual development. When one turns red, the email arrives before the customer calls, and the monitor name tells you at once which project it is.
Tips for several projects
- Name monitors consistently, for example “Project – purpose”, so the list stays readable.
- After larger changes, check that the monitored address still gives the right signal.
- Choose the check interval per project: shops and APIs where every minute counts more often, internal pages less often.
