1. Scope
This SLA applies to Alvo core components: workspace (alvo.energy/app), public API endpoints (`/api/*`), documentation and status pages. Excluded: third-party dependencies (ENTSO-E, OREE, WorkOS) which carry their providers' own SLAs.
2. Availability target (uptime)
Pilot target: **best-effort 99.5%** monthly, excluding planned maintenance. That's ~3.6 hours of downtime per month. GA target: **99.9%** (≤43 min/month).
Calculation: (available time) / (total time - planned maintenance) × 100%. /status shows current state, but a public uptime and postmortem archive is not live during pilot; reports are shared directly with the participant.
| Component | Pilot uptime | GA uptime | How measured |
|---|---|---|---|
| Workspace (/app) | 99.5% | 99.9% | synthetic alvo.energy/api/health probe at 5-min intervals |
| Public API (/api/*) | 99.5% | 99.9% | synthetic probes of key endpoints |
| ENTSO-E feed (coupling/multi-zone) | best-effort | best-effort | freshness status (fresh/stale/fallback) — depends on upstream |
| OREE feed (DAM/IDM prices) | best-effort | best-effort | depends on Market Operator publication schedule |
3. Planned maintenance windows
Routine maintenance (deploys, DB migrations, container rebuilds): Sundays, 03:00–05:00 UTC. Announced 48 hours in advance via email + in-product /account banner. Maintenance does not count toward downtime.
Emergency maintenance (security patches): no advance notice, with post-hoc notification.
4. Support response times
Channel: [email protected]. On business days (Mon-Fri, 9:00–18:00 EET) we meet the targets below. Weekends are best-effort.
| Tier | Description | First response | Resolution target |
|---|---|---|---|
| P1 — Critical | Platform down, data loss, security incident | 1 hour (business), 4 hours (weekend) | 8 hours |
| P2 — High | Core functionality broken for many tenants (forecast, AI brief, plan) | 4 hours | 1 business day |
| P3 — Normal | Single feature or UX question | 1 business day | 1 business week |
| P4 — Low | Docs, copy fixes, feature requests | 3 business days | Best-effort |
5. Incident communication
Public /status machine-reports current state and affected components. During pilot, operator-authored P1/P2 updates and ETAs are sent through the participant's agreed channel; a public incident editor is not live yet.
Postmortems are shared directly with pilot participants; a public /status archive is not live yet.
6. Exclusions
- Downtime caused by the customer (malformed API request, rate-limit overflow, revoked key)
- Downtime from third-party dependencies (ENTSO-E Transparency Platform, OREE, WorkOS, Resend, Paddle, LiqPay)
- Force majeure (war, natural disasters, regulator directives)
- Planned maintenance in the announced window
7. Service credits (GA only)
Credits do not apply during pilot (the service is free in demo + paid plans cancellable anytime). At production GA, service credits will accrue based on achieved uptime:
| Monthly uptime | Credit (% of monthly subscription) |
|---|---|
| ≥ 99.9% | 0% (SLA met) |
| 99.0% – 99.89% | 10% |
| 95.0% – 98.99% | 25% |
| < 95.0% | 50% |
Credits apply to the next invoice; not paid as cash. Request a credit at [email protected] within 30 days of the incident, referencing the postmortem.
8. Reporting
During pilot, monthly uptime metrics are shared directly with the participant; /status shows current rather than historical state. A public monthly archive and automated email reports are not live yet.
We negotiate custom uptime targets, RTO/RPO, support tiers, and service credits with enterprise customers under an MSA.
[email protected]