Procurement document

Service Level Agreement

The formal availability, recovery, response and remedy commitments. These commitments are incorporated by reference into every customer contract. Save as PDF (⌘P / Ctrl+P) for legal and procurement review.

Document version: 2026-05-14 · Latest version always at medicarehis.com/sla

§1 · Uptime

Availability commitment by tier

Monthly uptime measured as the percentage of minutes in a calendar month during which the system is reachable and serving requests on the public URL. Measured by independent third-party uptime monitor (UptimeRobot) plus our internal probes. Excludes scheduled maintenance windows declared at least 7 days in advance.

Tier Monthly uptime target Max downtime / month Service credit if missed
Starter99.5%~3h 39m5% of monthly fee for 99.0–99.5%; 10% below 99.0%
Hospital99.5%~3h 39m10% of monthly fee for 99.0–99.5%; 20% below 99.0%
Enterprise / Group99.9% (negotiable to 99.95% with multi-region)~43mCustom remedies negotiated per contract; typically 15–30% of monthly fee

99.95% requires multi-machine high availability + multi-region failover and is Stage 2 capability (unlocked when we move from single-machine bluegreen-equivalent to multi-machine Postgres). Enterprise customers requiring 99.95% today should request the per-hospital server deployment instead — that's a self-managed on-prem deployment with no internet-link dependency.

§2 · Disaster recovery

Recovery Point Objective & Recovery Time Objective

Metric Target Achieved today
RPO (data)< 1 hour~1 hour (hourly snapshots + tiered local retention + client-side encrypted off-host backup)
RPO (audit chain)< 5 minutes~60 seconds (mirror flush cadence to B2 Object Lock)
RTO< 1 hour~30 minutes per the documented disaster-recovery runbook
Backup retention (local)30 days48h hot + 28d daily + 30d age cap
Backup retention (off-host)≥ 90 days90-day hide + 30-day delete (120 days total) on B2
Audit-chain retention7 years7 years on B2 Object Lock Compliance mode (cannot be altered or deleted within retention window — even by us)

DR drills

Quarterly disaster-recovery drill exercises the full off-host restore path read-only without touching production: authenticates to both B2 buckets, downloads latest backup, decrypts client-side, extracts, verifies manifest SHA-256 hashes, runs the audit-chunk verifier. Drill output is committed to the governance directory as evidence; available to Enterprise customers on request.

§3 · Response times

Incident response matrix

When a customer reports an issue or our monitoring detects one, the response-time clock starts at the moment we receive notification. Different SLAs apply by severity:

Severity Definition Acknowledgement First fix shipped
Critical (P1)System unavailable; data loss or active breach in progress; clinical workflow blocked at scale1 hour, 24×77 days
High (P2)Major feature broken; one tenant affected; clinical safety concern4 hours, business hours30 days
Medium (P3)Bug with workaround; performance degradation; non-blocking display issue24 hours, business hours90 days
Low (P4)Cosmetic; feature request; suggestion7 daysBest effort; queued for roadmap

Enterprise tier negotiates tighter SLAs per contract (typically: P1 30-min ACK, 24-hour fix; P2 1-hour ACK, 7-day fix). A premium support uplift brings P1/P2 to 24×7 coverage with the Enterprise-tier targets.

§4 · Incident communication

How and when we tell you

§5 · Maintenance windows

Planned maintenance is predictable

Scheduled maintenance windows are announced at least 7 days in advance via email to your nominated technical contact and on the public status page.

§6 · Remedies

What happens if we miss

If we miss an uptime commitment in a calendar month, a service credit is automatically issued against your next invoice. Service credits are calculated as a percentage of the monthly subscription fee for that tier — see the table in §1.

Service credit cap

Force majeure

Outages caused by events outside our reasonable control (declared sub-processor outages, regional internet failures, regulatory action, natural disasters affecting our or our sub-processors' facilities) do not count against the uptime SLA. We will still communicate during force-majeure events per §4.

§7 · Measurement and reporting

How uptime is measured

Independent uptime measurement by UptimeRobot, probing the customer-facing URL every 5 minutes from multiple regions. Internal probes complement this with API-endpoint health checks every 60 seconds. Monthly uptime reports available on request for Hospital and Enterprise tier customers; quarterly for Starter tier.

Live component health, active incidents, planned maintenance, and 90-day uptime are public at medicarehis.com/status. Monthly per-customer reports remain available to Hospital and Enterprise customers on request.

Honest caveats

What we are NOT promising

Need negotiated terms for Enterprise tier?

Enterprise customers customise SLAs per contract — tighter response times, dedicated infrastructure, regional residency, custom remedies. Book a call with the founder to negotiate.

Companion documents: Procurement evidence pack · Brochure · Security whitepaper · Deployment architecture · Compliance roadmap · Support · Onboarding