Going live is the worst possible time to discover a connection problem. By the time the show has started, the audience is already watching the buffering. A short check beforehand is far less stressful than troubleshooting in front of a live audience, and it takes only a few minutes once it becomes routine.
This is a practical pre-flight check for anyone streaming from a venue, an office or a home studio. It is written for the moment just before you press go live, not for a deep network investigation.
A speed test is a starting point, not the whole picture
Most people check their connection with a one-off speed test and stop when it reads a healthy number. That number describes average download throughput, which matters for uploading a stream, but it does not describe how the connection behaves moment to moment.
Video and audio are sensitive to two things a speed test tends to hide:
- Latency, the round-trip time for data to reach the server and come back. When it rises, everything feels delayed and the encoder has less room to recover.
- Stability, or how much latency varies. A connection that averages 30 ms but regularly jumps to 300 ms will drop frames even though its average looks fine.
A single measurement is a photograph. A live stream needs something closer to a continuous readout.
What actually causes a stream to drop
In practice, drops usually come from one of a few causes:
- The link is shared. Someone else in the building starts a large upload and your available bandwidth falls below what the encoder needs.
- Wi-Fi is doing its job badly. Distance, walls, interference and a busy channel all add latency and packet loss.
- The connection is intermittent rather than slow. Brief outages of a second or two are enough to break a session, yet they are easy to miss when you are only looking at a speed number.
- The route to the ingest server changes. Congestion or a problem upstream can raise latency without affecting a nearby speed test.
Knowing which one you are dealing with changes what you do about it.
A five-minute pre-flight check
Run through this shortly before the show, ideally on the same machine and network you will actually stream from:
- Confirm the connection type. Wired Ethernet is still the safest choice for a fixed setup. If you must use Wi-Fi, make sure you are on the network you think you are.
- Measure latency, not just speed. Watch the round-trip time for a minute or two. You are looking for a low, steady figure rather than a single good reading.
- Watch for spikes. The useful question is whether the latency stays flat or jumps around. A flat 40 ms is better for streaming than an average of 40 ms that swings between 10 and 400.
- Test upload as well as download. Your stream travels upstream, and many connections have far less upload capacity than download.
- Do a short, real rehearsal. Encode ten minutes to the actual ingest server at the settings you plan to use. This is the only test that reflects the whole path.
If you want a continuous read of the first three without installing anything, Conntivity measures HTTPS round-trip time every second in the browser, tracks it over time and flags the moments when the connection actually drops. It is a quick way to see whether the link is steady or quietly misbehaving before you commit to a show.
What to watch while you are live
You cannot watch a graph for the whole show, but a few things are worth glancing at between segments:
- Whether the latency readout is drifting upward, which often precedes trouble.
- Whether the connection has briefly dropped, even if it recovered.
- The encoder's own health, such as dropped frames or a falling bitrate.
If the numbers start to move, you have time to lower the bitrate or switch to a backup path before the audience notices.
If the connection does drop
When something fails mid-stream, resist the urge to change several things at once. Work through it in order:
- Check whether the whole connection is down or just the stream. If other services are reachable, the problem is more likely the ingest server or the encoder.
- Note whether the drop was a clean outage or a period of high latency. They point to different causes.
- Fall back deliberately. A rehearsed backup, such as a second connection or a lower bitrate, is worth more than an improvised one.
Browser checks and deeper tools
For a fast, no-install check, a browser-based monitor is enough to answer the common question: is the link steady right now? It measures the same kind of round-trip time your apps experience, which is often more relevant than a synthetic ping.
For native ICMP ping, traceroute and a fuller picture of the path, a dedicated tool is the better fit. MyEvent Labs builds netdx for exactly that kind of path analysis, and Conntivity for the quick in-browser pre-flight. Used together, they cover the check before the show and the investigation when something looks wrong.
The habit that matters most is simply having one. A few minutes of checking before you go live is the cheapest insurance a stream has.