Service Status & Uptime¶
FinMind publishes a public, real-time status page:
It shows the current API status, uptime over the past 24 hours and 90 days, and a history of past incidents.
This uptime is also the measurement basis for the FinMind enterprise plan SLA — enterprise service levels are assessed with the same per-minute algorithm, kept transparent and consistent with what the status page shows.
How uptime is calculated¶
Uptime is measured from real API traffic, minute by minute. The full formula chain is:
1. Valid requests
- A server-side error means an HTTP 5xx response, plus a connection dropped by a server-side timeout (HTTP 499 — usually caused by the server being busy or slow, so it is treated as a server-side problem).
- Ordinary client errors (e.g.
400/403/404/429, excluding 499) are not counted as valid requests and do not affect uptime.
2. Per-minute error rate
- A minute with traffic and an error rate > 5% is counted as down.
- Minutes with no traffic are not counted toward uptime.
3. Uptime
- The 24-hour and 90-day figures aggregate these per-minute results; the 90-day and monthly numbers are weighted by minutes with traffic.
- Scheduled maintenance windows may be excluded.
Status levels¶
When a time window has traffic, it is classified into four levels by that window's uptime:
| Status | Uptime threshold |
|---|---|
| Operational | 100% (traffic present, no down minutes) |
| Degraded Performance | ≥ 99% (some down minutes, but uptime still ≥ 99%) |
| Partial Outage | 95% – 99% (95% ≤ uptime < 99%) |
| Major Outage | < 95% |
| No data | No traffic in the period |
Tip
The same "How uptime is calculated" note is also available at the bottom of the status page itself.