How Hosting Uptime and SLA Measurement Work

A current UP state, a measured availability percentage and a contractual SLA commitment are three different facts.

Evidence firstCollect current state before changing it

Rollback awareKeep a verified route back

No credentialsNever paste passwords into a public checker

Current status answers only one question

A successful check shows that the tested endpoint worked at that moment. It does not reconstruct earlier outages and it does not prove that every application path is healthy.

Availability needs a complete window

A percentage requires a defined start and end, regular samples, a success rule and a policy for monitoring gaps. Publishing a percentage before a full window exists creates false precision.

The SLA is a contractual target

The FastVPS SLA level applies only when the final order states it. The covered service, eligible downtime, exclusions, incident process and service-credit calculation are defined in the SLA document.

Measure the service users actually need

ICMP reachability is useful but insufficient for a website or mail platform. Public checks should test DNS, TCP, TLS and the relevant protocol or application response.

Keep maintenance and customer faults distinct

Scheduled maintenance, customer application faults and third-party dependencies may be treated differently from platform downtime. The measurement policy must say so before an incident.