A 30-second interval sounds like a small setting. It changes the amount of work the monitoring system does all day, including when every service is healthy. For a six-person company, that background work deserves as much attention as a new feature.
One check running every 30 seconds produces 2,880 scheduled rounds a day. At a three-minute interval, the same check produces 480. That is a sixfold difference before we account for regional requests, confirmation attempts or retries. The interval is part of why Free has a different allowance from our paid plans.
Requests are only part of the bill
Each request needs a worker, a connection and a result record. We also keep timing information so you can compare Frankfurt, London, Virginia, Oregon, Singapore and Sydney. A short timeout does not make the work disappear; it means the worker has spent time waiting for a response that never arrived.
Confirmation adds requests during failures. That is exactly when many endpoints may be failing together. We need room for that extra work instead of sizing the system only for a quiet afternoon. Otherwise the monitoring queue could slow down when teams most need it.
Storing results has its own cost. Business keeps 13 months of history, and the charts need to stay useful as that history grows. We summarize data for longer chart ranges while retaining the records covered by the plan. A CSV export lets a team do a different analysis without asking the chart to answer every question.
Keeping the workload predictable
We spread scheduled work across the interval so checks do not all start at the same second. Regional workers report their queue delay as well as request duration. Those are different numbers: a slow queue points to our scheduling, while a slow response may point to the service being checked.
We also limit how much data a check downloads. Monitoring a predictable endpoint is cheaper for both sides than repeatedly fetching a large page with no useful health signal. When someone asks which URL to use, we suggest a small response that still exercises the part of their service they care about.
The monthly price pays for more than the request loop. It covers result storage, public status pages, alert routing and time spent investigating our own failures. SMS allowances are visible because message delivery adds a separate usage cost.
We do not want a low headline price followed by unexplained charges. Check slots and message allowances put a boundary around the work. When you reach one, the account shows the limit and the available options before you spend more.
