How Covered.ir measures
This page explains where the dashboard's numbers come from and what they actually claim — and, just as importantly, what they do not.
What is measured
For each defined route between two points (a datacentre, an operator, or a city cluster), round-trip latency and the percentage of lost packets are measured continuously. The measurement is of the route itself, not of any particular website or service.
How it is measured
Monitoring agents installed at the endpoints run ICMP and TCP checks and report the results to a central Zabbix system. Covered.ir reads and presents that same data; it does not generate or estimate any figure of its own.
Update cadence
Route status is refreshed every few tens of seconds and pushed to the dashboard over a WebSocket, with no page reload. The time of the most recent measurement is always shown in the header.
What the states mean
- Healthy
- The route is operating within its normal range: negligible packet loss and latency close to that route's own baseline.
- Degraded
- The route is up but its quality has dropped — measurable packet loss, or latency meaningfully above its own baseline.
- Incident
- A serious, confirmed disruption on the route; services depending on it are very likely affected.
- Down
- No response is being received from the route at all.
- Unknown
- No data is available. "We cannot tell" is deliberately never shown as green.
Comparison against baseline
Whether a route is healthy is a relative judgement: Tehran–Frankfurt at 95 ms is healthy, while the same figure on Tehran–Isfahan would be a serious fault. Each route is therefore judged against its own historical baseline rather than one global threshold.
Limitations
This dashboard shows the state of the routes it monitors, not the state of the country's entire internet. If a route is not listed here, no claim is made about it. Measurements are also taken from our endpoints' point of view; a home user on a different operator may well experience something else.
Current data status
Until the connection to the monitoring system is complete, the dashboard runs on demonstration data. That data always shows a healthy state so that an invented outage is never mistaken for a real one, and it is labelled explicitly in the page footer.