1. Uptime commitment

Nodara guarantees 99.995% monthly uptime per instance, measured from the network fabric to the instance's tenant network. That works out to roughly 2 minutes 11 seconds of allowed unavailability per 30-day billing cycle.

What counts as downtime

• Instance is unreachable via its public IPv4 for > 60 seconds while our monitoring reports the tenant network as healthy.
• Console access unavailable while the instance shows as running.
• Snapshot restore fails on a snapshot marked as verified.

What does not count

• Scheduled maintenance announced ≥ 72 hours in advance.
• Failures caused by the customer (misconfigured firewall, OS crash, resource exhaustion inside the guest).
• Force majeure events (regional power grid failure, undersea cable cuts, government orders).

2. Credit schedule

Credits are applied automatically to the next invoice — you never need to file a ticket. Credits stack across concurrent incidents up to a monthly cap of 100% of the affected instance's monthly rate.

3. Response commitments

Support response times are measured from ticket creation to the first substantive human reply — not an autoresponder.

• P1 (production down): 15 minutes, 24/7.
• P2 (degraded): 1 hour during business hours (00:00 – 20:00 UTC).
• P3 (general): 4 business hours.

4. Data durability

NVMe volumes are triple-replicated within a region across independent power domains. Weekly snapshots are stored in a second geographic zone at no extra cost. We target 11 nines (99.999999999%) of annual durability on snapshotted data.

5. Filing a claim

Credits are calculated and applied automatically. If you believe your instance experienced qualifying downtime not reflected in your next invoice, email sla@nodara.host with the instance ID and observed outage window. We reply within one business day.