Service Level Agreement (SLA)

Last updated: February 2026

Uptime Commitments

Plan Tier Monthly Uptime Target Max Downtime/Month
Starter (non-HA) Best effort (~99.5%) ~3.6 hours
Professional HA 99.9% ~43 minutes
Business HA 99.95% ~22 minutes
Enterprise HA 99.95% ~22 minutes

HA plans use multi-node redundancy: Docker Swarm auto-reschedules on node failure, MariaDB Galera provides 3-node synchronous replication, GlusterFS provides 3-way replicated storage.

How Uptime Is Measured

  • Monitoring tool: Prometheus Blackbox Exporter polling each site every 30 seconds.
  • Downtime defined as: Site returning non-2xx HTTP status for 2 or more consecutive minutes as measured by our monitoring infrastructure.
  • Uptime calculation: ((total_minutes - downtime_minutes) / total_minutes) × 100

What Counts as Downtime

  • Complete site inaccessibility (connection refused, timeout)
  • Persistent 5xx errors from our infrastructure (not customer code)
  • SSL certificate expiry caused by our systems
  • Database unavailability affecting site operation

Exclusions (Not Counted as Downtime)

  • Customer-caused issues: Bad plugin code, PHP fatal errors, theme bugs, .htaccess misconfiguration
  • DNS propagation: Delays at the customer's registrar or DNS provider
  • Scheduled maintenance: Announced at least 24 hours in advance via email; typically performed during low-traffic hours (02:00-06:00 UTC)
  • Force majeure: Natural disasters, acts of war, government actions, widespread internet outages
  • Third-party services: Stripe, SendGrid, Cloudflare, or other external service outages
  • DDoS attacks: Beyond the capacity of our mitigation infrastructure
  • Customer account suspension: Due to non-payment, TOS violation, or abuse

How We Achieve It

Layer Redundancy Failover
Web/App Docker Swarm (2+ replicas on HA plans) Auto-reschedule on node failure (~30s)
Database MariaDB Galera (3-node synchronous replication) Automatic via Galera cluster consensus
Storage GlusterFS (3-way replicated across all nodes) Transparent to application
SSL/Proxy Caddy with automatic Let's Encrypt Certificate auto-renewal
Monitoring Prometheus + Blackbox + Grafana Automated alerting

Remedies

If verified downtime exceeds the committed uptime target in a calendar month:

Excess Downtime Service Credit
Up to 1 hour beyond SLA 5% of monthly fee
1-4 hours beyond SLA 10% of monthly fee
4-12 hours beyond SLA 25% of monthly fee
Over 12 hours beyond SLA 50% of monthly fee

Maximum credit: 50% of the monthly fee for the affected service in the affected month.

How to claim: Contact support within 30 days of the incident. Include the affected site domain and the approximate time window. We will verify against our monitoring data.

Starter (Non-HA) Plan Note

Starter plans run a single WordPress replica. If the node hosting the site goes down, the site is unavailable until Swarm reschedules the container (typically 1-5 minutes) or, in the case of a full node failure, until the node recovers or the site is manually redeployed. No uptime SLA is guaranteed for non-HA plans; the "best effort" target of ~99.5% reflects historical platform availability.

Upgrade path: Customers requiring guaranteed uptime should upgrade to a Professional or Business HA plan.

Changes to This SLA

We may update this SLA from time to time. Changes will be announced via email to active subscribers at least 30 days before taking effect. The current version is always available on this page.