Service Level Agreement

Effective 10 January 2025. This SLA applies to every RDPForge instance on an active, paid service.

The commitment

We commit to 99.9% availability per instance per calendar month, which allows for 43 minutes and 12 seconds of unavailability. Measured fleet-wide uptime has been above this every year since 2021 — 99.98% in 2025.

How availability is measured

An instance is unavailable when it fails ICMP and TCP reachability checks from at least three of our four external probes for two consecutive one-minute intervals, and the cause lies within our infrastructure. Probes are hosted outside our network on independent providers, deliberately, so we are not marking our own homework.

Partial degradation — elevated latency or packet loss above 2% sustained for more than 15 minutes — counts as unavailability at 25% weighting.

Credit schedule

Monthly availability	Downtime in a 30-day month	Credit
99.9% – 100%	up to 43 m	None
99.5% – 99.89%	43 m – 3 h 36 m	10% of the monthly fee
99.0% – 99.49%	3 h 36 m – 7 h 12 m	25%
95.0% – 98.99%	7 h 12 m – 36 h	50%
Below 95.0%	more than 36 h	100% of the monthly fee

Credits apply to the affected instance only, are calculated on its effective monthly rate, and are capped at 100% of one month's fee per calendar month.

Automatic application

Since January 2025 credits are applied automatically. Our monitoring computes per-instance availability at the end of each month and issues the credit against your next invoice with a line item naming the incident. You do not need to notice, calculate or claim anything.

Exclusions

The commitment does not cover unavailability caused by:

• Your own configuration — a firewall rule, a failed service, a full disk, a broken kernel upgrade, or an OS that will not boot after your change.
• Scheduled maintenance announced at least 72 hours in advance. We aim for under 4 hours of announced maintenance per instance per year and have averaged 51 minutes since 2022.
• Suspension for non-payment or AUP violation.
• Attacks against your instance that exceed our mitigation capacity, or that require a temporary null-route to protect other customers. In practice this has happened twice since 2020.
• Force majeure — natural disaster, war, government action, or a facility-level failure entirely outside our control and outside the facility operator's own SLA.

Network latency commitment

We commit to sub-1% packet loss and stable latency within each region's normal baseline. Sustained deviation caused by our network or our transit selection is treated as degradation under the schedule above. Latency between your location and our datacenter over networks we do not control is outside this commitment — which is why every region publishes a test IP so you can measure it before buying.

Sole remedy

Service credits are the sole and exclusive remedy for unavailability. This does not limit your rights under the refund policy or any right that cannot lawfully be excluded.