playback stall after

[M1Z8N]

Summary

Seeing occasional stalls where RTMP publish continues (encoder shows OK) but downstream playback hangs or buffers indefinitely after ~3–5 minutes.

NOTE: This is a placeholder issue to discuss stability/telemetry needs. Feel free to close if not actionable yet.

Expected

Continuous publish with uninterrupted playback.

Actual

Playback buffers/hangs; restarting the player resumes for another few minutes.

Repro (placeholder)

1. Start local publisher (e.g., ffmpeg -re -stream_loop -1 -i sample.mp4 -c copy -f flv rtmp://localhost/live/test)
2. Open player (VLC/ffplay) to rtmp://<host>/live/test
3. Observe buffering/hang after ~3–5 mins.

Env

• RapidRTMP: <commit/branch if known>
• OS: macOS 14 / Ubuntu 22.04
• CPU: x86_64
• Publisher: ffmpeg 6.x (H.264 + AAC)
• Player: ffplay/VLC/OBS

Logs / ideas

• Would be helpful to expose server-side metrics (bitrate in/out, queue depth, dropped packets).
• If there’s a session-level keepalive/timeout, maybe we can surface it in logs.

Proposed next steps

• Add minimal telemetry: per-stream bitrate, client count, and rolling buffer stats.
• Optional: expose /healthz and /metrics (Prometheus) for quick triage.

[hunter-nguyen]

lgtm

[M1Z8N]

lgtm