How do I add a backup stream for redundancy?
For a broadcast that matters — a paid event, a service, a tournament final — a single dropped connection shouldn’t end the show. Redundancy means having a second feed ready to take over the instant the first one falters. Here’s how backup streams and automatic failover work.
dropped out
2nd encoder
Two kinds of redundancy
There are two failure points to protect: the outbound relay to each platform, and the inbound feed from your encoder. Supervising the outbound relays — restarting any that drop — handles platform-side hiccups. A backup ingest feed handles the bigger risk: your source itself going down.
How backup ingest failover works
You publish your primary feed as usual, and a second encoder publishes a parallel feed to a backup ingest endpoint. While the primary is healthy, it’s what gets relayed. If the primary drops, the relay automatically falls back to the backup feed — so your destinations keep receiving video instead of going dark.
The two encoders should be as independent as you can make them: ideally different machines on different networks (for example, a wired primary and a bonded-cellular backup), so a single point of failure can’t take out both at once.
Don’t forget the boring stuff
Redundancy is only as good as its weakest link. Use wired connections where you can, give your encoder bitrate headroom, and test the failover before the real event — pull the primary feed and confirm the backup takes over the way you expect.
- Primary + backup encoders on separate machines and networks.
- A relay that supervises outbound destinations and restarts failures.
- A rehearsal: drop the primary and watch the backup take over.
Frequently asked questions
What is backup ingest / stream failover?
A second encoder publishes a parallel feed to a backup endpoint; if your primary source drops, the relay switches to the backup so your destinations stay live.
Does failover prevent every kind of outage?
It protects against a dropped source and failed outbound relays. It can’t fix a platform-side outage or your viewers’ connections — for the parts we control, it keeps you on air.
Should the backup encoder be on a different network?
Ideally yes — a separate machine and connection (e.g. wired primary, cellular backup) means one failure can’t take out both feeds.