# Streamloop status

> Live status of Streamloop, which streams uploaded videos 24/7 to YouTube, Twitch and any RTMP destination. Generated 2026-10-22T22:41:31.113Z from the checks behind https://status.streamloop.app/. Overall: up.

- Status page: https://status.streamloop.app/
- Incidents and maintenance (RSS): https://status.streamloop.app/feed
- Components as JSON: https://status.streamloop.app/index.json

## Components

### Dashboard & API
- Dashboard: up; checked by Better Stack every 3 min.
- Public API: up; checked by Better Stack every 3 min.
- Sign-in: up; checked by Better Stack every 3 min.
- Authenticated API calls: up; last report 2026-10-22T22:41:27Z (just now): passed.
- Website: up; checked by Better Stack every 3 min.

### Stream scheduling
- Scheduled starts and stops: up; last report 2026-10-22T22:41:26Z (just now): passed.
- Stream launches: up; last report 2026-10-22T21:54:16Z (47 min ago): passed.

### Live streaming
- Stream delivery: up; last report 2026-10-22T22:41:26Z (just now): passed.

### Video encoding
- Video processing: up; last report 2026-10-22T22:41:27Z (just now): passed.

### Video storage & uploads
- Uploads and storage: up; last report 2026-10-22T22:33:29Z (8 min ago): passed.

## Metrics

- All streams, time to go live (24 h to 2026-10-22T22:38:28Z): median 43 s, 95th percentile 261 s; 98.5% of launches went live (133 live).
- Test stream from a fresh upload (2026-10-22T21:54:16Z): live 27.9 s after Start; the upload took 0.7 s.
- API (15 min to 2026-10-22T22:38:29Z): 100.00% of 693 requests without a server error; time to first byte median 138 ms, 95th percentile 328 ms.

## Incidents and maintenance (last 7 days)

- None reported.

## How this is monitored

Two independent sources feed this page, and a third adds metrics.

1. Better Stack checks the public endpoints itself, every 3 minutes from 4 regions (US, EU, Asia, Australia):
   - Dashboard: https://streamloop.app/auth/login must answer 200.
   - Public API: https://api.streamloop.app/v1/status must answer with "operational".
   - Sign-in: https://auth.streamloop.app/api/auth/jwks must answer with its signing keys.
   - Website: https://streamloop.app must answer without an error.

2. A prober on a separate server, outside Streamloop's Kubernetes clusters, reports to Better Stack heartbeats.
   Silence means down: a heartbeat that is not pinged within its period plus grace goes down, so the prober
   failing never reads as healthy. A check that finds a problem reports a failure at once.
   - Stream delivery, Scheduled starts and stops, Video processing: every minute the prober reads the findings
     of Streamloop's sentinel, a job inside the cluster that runs every 5 minutes and checks the live streams,
     their schedules and their encodes. With no readable sentinel run in the last 30 minutes the prober stays
     silent, and these go down after about 31 minutes.
     - Stream delivery is down when at least 3 streams, and at least 20% of the streams that should be sending,
       send nothing to their destinations; degraded at 3 streams and 2%, when many streams are torn down at
       once, or when 3 or more streams are stuck failing.
     - Scheduled starts and stops is down when 10 or more scheduled starts, stops or repeats are overdue;
       degraded at 3, or when 3 or more launches or stops have been stuck for over 30 minutes.
     - Video processing is degraded, never down (these signals cannot tell the two apart), when encodes fail
       across a 3-hour cohort, when 3 or more live streams lose their encoded tail, or when 3 or more launches
       are stuck preparing for over 30 minutes.
   - Authenticated API calls: every minute, GET /v1/me with a dedicated API key. Down after 5 minutes of silence.
   - Uploads and storage: every 10 minutes, create an upload, PUT a file, complete it, download it back,
     compare its SHA-256 and delete it. Down after 25 minutes of silence.
   - Stream launches: every hour a 30-second video nobody has seen is uploaded, put into a test stream and
     started; the launch counts once its frames decode at Streamloop's own RTMP receiver. Down when a launch
     fails or is not live within 15 minutes, or after 90 minutes of silence.

3. Metrics, computed every 15 minutes from Streamloop's observability store:
   - All streams, time to go live: median and 95th percentile from pressing Start to live, over the previous
     24 hours of customer launches. Failures on the customer's side (wrong stream key, no credits, stopped
     before going live) are left out.
   - API success rate and response time: requests to api.streamloop.app as Cloudflare's edge saw them over
     15 minutes; the share answered without a server error, and the time to first byte.

Not monitored yet: a canary stream that runs continuously, with frame-level checks every minute.