# TrackMyVend Service Level Agreement (SLA) **Effective date:** 2026-04-21 · **Revision:** v1.0 This document defines the availability commitment, response times, and remedy terms TrackMyVend offers its paying customers. A plain-text copy is served at and linked from the public status page at . --- ## 1. Service availability TrackMyVend commits to **99.9 % monthly uptime** for the production web application and API, measured as: ``` uptime_% = (total_minutes - unplanned_downtime_minutes) / total_minutes * 100 ``` Where a month of 30 days has 43 200 minutes, **99.9 % allows up to 43 minutes of unplanned downtime per month.** ### What counts as "available" The system is considered available when **all** of the following are true, as observed by our external monitors: - The public landing (`/`) returns HTTP 200 (or a 302 to a locale subdirectory). - The authenticated app (`/dashboard.html`) returns HTTP 200 to a valid session. - `/api/health` returns HTTP 200. - Round-trip latency from the monitor nodes in North America and Latin America is under 3 000 ms. ### What does NOT count as downtime - Scheduled maintenance announced at least 48 h in advance. - Outages of external dependencies explicitly outside the app boundary (Mapbox, the customer's own ISP, etc.). We'll disclose these in status-page notes but won't count them against the SLA. - Force majeure events. ## 2. Availability tiers by plan | Plan | Monthly uptime target | Response time (P0) | Credit cap per month | |---|---|---|---| | Starter | 99.5 % | 4 h business-hour | 10 % of that month's fee | | Pro | **99.9 %** | 1 h business-hour | 25 % of that month's fee | | Enterprise | 99.95 % | 15 min 24/7 | 50 % of that month's fee | ## 3. Incident severity | Level | Definition | Example | |---|---|---| | **P0 — Critical** | Service unreachable for > 1 % of users OR data loss risk | Supabase outage; auth broken across the board | | **P1 — High** | A core workflow is broken but sessions still work | Dashboard KPIs show wrong totals | | **P2 — Medium** | A secondary feature is broken; workaround exists | Map export PDF fails | | **P3 — Low** | Cosmetic / non-blocking | A chart label wraps oddly | Response SLAs (from first customer report to first human response): | Severity | Starter | Pro | Enterprise | |---|---|---|---| | P0 | 4 h (business hrs) | 1 h (business hrs) | **15 min (24/7)** | | P1 | 1 business day | 4 h (business hrs) | 1 h (business hrs) | | P2 | 2 business days | 1 business day | 4 h (business hrs) | | P3 | 5 business days | 2 business days | 1 business day | Business hours = 09:00–18:00 CST Monday–Friday, excluding Mexican public holidays. ## 4. Service credits (how we compensate downtime) If we miss the monthly uptime target on a paid plan, the customer can request a service credit: | Actual monthly uptime | Credit (% of monthly fee) | |---|---| | 99.0 % – 99.9 % | 10 % | | 95.0 % – 99.0 % | 25 % | | Below 95.0 % | 50 % (plan cap) | Credits are applied to the next invoice. Customers must request the credit in writing (billing@trackmyvend.com) within 30 days of the incident. ## 5. How we measure availability - External synthetic monitors (BetterStack / UptimeRobot) hit `/api/health` from **at least 3 geographic regions** every 60 seconds. - Two consecutive failures are required before we mark an incident (reduces false positives from transient network blips). - Monthly uptime is posted on within 5 business days of month-end. ## 6. Planned maintenance windows When we need to push a breaking change: - Announced at least **48 h in advance** on the status page, by email to active admins, and on our Slack Connect channels (if applicable). - Scheduled between **02:00 – 06:00 CST** Tuesday or Thursday whenever possible. - Never during Monterrey / São Paulo / Buenos Aires business hours if avoidable. ## 7. Exclusions The SLA does not apply to: - Beta features clearly labeled as such in the product. - Sandbox / staging environments. - Customer-owned hardware (the Nanolink IoT devices — they're covered by a separate hardware warranty). - Downstream dependencies entirely outside our control (Mapbox tile outage, Supabase regional outage, etc.) — we will still publish these on the status page but they do not count against the uptime target. ## 8. Contact - **Incident reporting:** soporte@trackmyvend.com - **Status page:** - **Billing / credits:** billing@trackmyvend.com - **Enterprise escalation:** see your CSM's signature line --- *TrackMyVend reserves the right to update this SLA with 30 days' advance notice. The current authoritative version lives at . Previous revisions are kept in the Git history of the website.*