Each restart cuts the picture for every viewer of the channel. The panel counts restarts and keeps a history, which often points at the cause.
Checks
- Count the restarts. On Streams → Manage Streams, the Restarts column counts the channel's recent restarts. 0 is calm, up to 3 is worth a look, and more means the channel keeps failing.
- Read the history. In the row menu (Options), open Restart history. It lists each restart, newest first, with the source the channel was on.
- Restarts at a steady interval, for example every hour: the provider probably ends connections after a set time.
- The same source again and again: check that source's login, and its connections at the provider.
- Check the provider's connections. Open Streams → Providers. Connections shows how many of the login's connections are in use, counting every player on it. When it is full, the provider refuses new connections or drops old ones. To free some, put little-watched channels on demand (see On demand or always on), or ask the provider for more.
- Add a backup source, from another provider if you can: see Backup sources and failover.
- Check the server. If many channels on one server restart together, open Servers → Server Stats: it shows the server's processor, memory and traffic over time. Converting takes about one processor core per HD channel: see Transcoding profiles.
- Try the other engine. On the channel's Settings step, open Engine. Leave it on GStreamer (recommended) unless a source only plays with FFmpeg.
How restarts work
- When the source in use fails, the server moves to the next source in the list.
- If every source fails, the channel shows Error. The server keeps trying the sources in turn, waiting a little longer each time (5 minutes at most), and the channel comes back by itself when a source works again.
- An always-on channel on a backup goes back to its main source once the main has answered every check for an hour, or when you restart or edit the channel.
Updated 12 November 2026