Service Level Agreement (SLA) v1.0

Platform v4.1 · Issue Date: 2026-05-06 · Ref: SATCLIMATE-SLA-2026-001

Confidential · Institutional Document

Engine Level Agreement

SatClimate Intelligence Platform — Carbonaa SA

Version: v1.0

Platform: v4.1

Issue Date: 2026-05-06

Doc Ref: SATCLIMATE-SLA-2026-001

Governing Law: Swiss Law · Geneva

99.9%

Enterprise Uptime Target

300ms

Enterprise P50 Latency

15 min

P1 Incident Response

1. SLA Targets by Subscription Plan

Explorer

CHF 2,500/month

Uptime Target

99.0%

≤ 7.3 hours/month

API Latency P50

< 800 ms

P99: < 3,000 ms

Signal Freshness

T+10 min (cached; up to T+24h fallback)

Support Response

5 business days (email only)

Incident Alarm

Best effort — status page

SLA Credits

No SLA credits at Explorer plan

Professional

CHF 8,800–CHF 10,000/month

Uptime Target

99.5%

≤ 3.6 hours/month

API Latency P50

< 500 ms

P99: < 2,000 ms

Signal Freshness

T+5 min (live; up to T+1h on fallback)

Support Response

2 business days (email + ticketing)

Incident Alarm

≤ 30 min for P1 incidents

SLA Credits

5% monthly credit per 0.5% below target (cap 15%)

Enterprise

CHF 40,000/month

Uptime Target

99.9%

≤ 43 minutes/month

API Latency P50

< 300 ms

P99: < 1,000 ms

Signal Freshness

T+5 min (live; fallback capped at T+30m)

Support Response

4 business hours (email + phone + dedicated)

Incident Alarm

≤ 15 min for P1 incidents

SLA Credits

10% monthly credit per 0.1% below target (cap 30%)

2. Incident Severity Classification

P1CriticalResponse: 15 min (Ent) / 30 min (Pro)Resolution: 4 hours

Complete API unavailability or >50% of endpoints returning errors

P2MajorResponse: 1 business hourResolution: 24 hours

Degraded performance (P99 latency >2× SLA target) or signal freshness >2× target

P3MinorResponse: 1 business dayResolution: 5 business days

Single endpoint degraded; non-critical feature unavailable; cosmetic issues

P4AdvisoryResponse: 2 business daysResolution: None required

Informational; no engine impact; planned maintenance alarms

3. Uptime Measurement Methodology

Measurement period: Calendar month (00:00 UTC on the 1st to 23:59 UTC on the last day).

Uptime definition: The percentage of minutes in a calendar month during which the Signals API and Environmental Data API return valid responses (HTTP 200 with non-error payload) for at least 95% of requests sampled at 5-minute intervals.

Downtime definition: Any period of ≥ 5 consecutive minutes where the uptime condition above is not met. Partial minutes are counted as full minutes.

Measurement source: Internal synthetic scanning probes. External scanning tools (e.g., Pingdom, UptimeRobot) may be used by customers as supplementary evidence but are not the primary measurement source for credit calculations.

Credit request procedure: Submit credit requests within 10 business days of the end of the affected month to amin@carbonaa.net with logs and timestamps. Credits are applied to the following billing cycle.

4. SLA Exclusions

The following events are excluded from uptime calculations and do not trigger SLA credits:

Scheduled maintenance windows (minimum 48 hours advance notice via status page)

Force majeure events (natural disasters, war, government action, pandemic)

Satellite operator outages beyond SatClimate control (NASA FIRMS, ESA Copernicus, NOAA)

Third-party infrastructure failures (cloud provider, CDN, DNS) not attributable to SatClimate

Client-caused outages (malformed requests, quota exhaustion, API key compromise)

Sandbox and free-plan access — no SLA applies

Beta or preview features explicitly labelled as such

Intermittent issues not reproducible or lasting < 5 minutes

5. Recovery Objectives (RTO & RPO)

These recovery objectives apply to the SatClimate API (contextualSignalsAPI + satclimateDataAPI) as a stateless signal delivery service.

RTO — Recovery Time Objective

4 hours

(P1 / Critical — complete service unavailability)

• Explorer product: Best efforts (no RTO guarantee)

• Professional product: 4 hours (P1 incidents)

• Enterprise product: 2 hours (P1 incidents)

• P2 / Major degradation: 8 hours (Pro), 4 hours (Ent)

RPO — Recovery Point Objective

0 minutes

(Signal delivery — stateless by design)

• Signal data: RPO = 0 — signals generated on-demand; no state lost during outage.

• Account data: RPO ≤ 24 hours (Base44 daily backups).

• AuditLog records: RPO ≤ 24 hours — not retroactively reconstructed.

6. Planned Maintenance

Standard maintenance window: Saturdays 02:00–06:00 UTC (4-hour window). Notified ≥ 48 hours in advance via status page at /SatClimateAPIStatus.

Emergency maintenance: May occur without advance notice in cases of active security incidents, data integrity risks, or third-party infrastructure failures.

Planned maintenance exclusion: Maintenance windows notified ≥ 48 hours in advance are excluded from uptime calculations.

SATCLIMATE-SLA-2026-001 · SLA v1.0 · Platform v4.1 · SatClimate / Carbonaa SA · Geneva, Switzerland · amin@carbonaa.net

This SLA is incorporated into and subject to the SatClimate Terms of Engine v3.0. Governing law: Swiss Law, Canton of Geneva. Not for public distribution without authorization from Carbonaa SA.

base44
Edit with Base44