Service Level Agreement

Last updated 2026-08-27

99.95%

Target Monthly Uptime Percentage for the product API — a direct pass-through of Google Cloud Run's own published SLA, not a number set independently of it.

Entropa's application code is test-covered and gated on every commit — the realistic source of any downtime is the underlying infrastructure, not application bugs slipping through untested. So the commitment that actually governs is the one Google Cloud already makes for the services Entropa runs on:

Cloud Run (the API itself) — 99.95% in most regions. Current terms →
Firestore (durable storage) — its own published SLA. Current terms →
Pub/Sub (the ingest queue) — its own published SLA. Current terms →

If Google's own terms for any of these change, this page is updated to match.

What counts as downtime

A minute counts as downtime only if a majority of well-formed POST /api/tx requests during that minute receive an HTTP 5xx response or time out. Downtime caused by your own integration (malformed requests, an expired or revoked API key, exceeding a fair-use rate limit) does not count.

Service credits

Monthly uptime	Credit
< 99.95% but ≥ 99.0%	10% of that month's charges
< 99.0% but ≥ 95.0%	25% of that month's charges
< 95.0%	50% of that month's charges

To request a credit, email aimozart@entropa.space within 30 days of the incident. Credits apply to future invoices, not cash payouts.

What this doesn't cover

Scheduled maintenance announced at least 24 hours in advance.
Any exclusion Google's own Cloud Run/Firestore/Pub-Sub SLAs carry (force majeure, factors outside Google's reasonable control) — Entropa inherits these since it cannot be more available than the infrastructure it runs on.
The marketing website (entropa.space) or the Scryon demo explorer — this SLA covers the product API only.