Four facts, and they take three minutes to collect. Scope: does the fault hit one channel, one whole category, or everything? Timestamp: the date and clock time, with your time zone, because server-side events are found by time. Environment: which player, which version, on which device, and whether anything updated recently. Recovery: switch away from the channel and straight back, because if the picture returns immediately the session stalled and bandwidth was never the cause. A message carrying those four gets routed to someone who can look at a server. A message saying nothing plays gets the restart script, because there is nothing else to send.

1. 1Scope: one channel, one group, or everything
2. 2Timestamp with your time zone
3. 3Player, version, device, and any recent update
4. 4Whether it recovers on switch-away-and-back

Scope decides most of it. One dead channel is a source problem at that stream, and nobody's settings fix it. One whole group failing while others play points server-side, at a feed or a node. Everything failing at once points at the line, the player, or an expired session. Then run two cheap tests. Put the same login into a second player on a different device: if it plays there, the fault is the app, which matters because community reports keep describing players that broke the day an operating system or app updated. And compare 2pm with 8pm on the same channel, because a fault that exists only at peak is capacity rather than configuration.

Write it as a report rather than a complaint. One line of symptom, then the four facts, then what you already tried, in that order. Name the channel or group exactly instead of describing it. Give the time with your time zone. State the player and version, the device, and whether the same login works elsewhere. For a billing question, give the account email and the payment date. Never put full card details in a support message, and a service that asks for them in chat has told you something important. Then ask for one specific outcome: a fix, an explanation, or a refund inside the window.

Peak-hour capacity, mainly. If a group is clean at 2pm and unwatchable at 8pm, that is server load, and the honest answer from any service is a capacity answer rather than a settings answer. Published uptime is where that belongs: 99.99% works out to roughly 53 minutes a year, and a service willing to state a figure has given you something to hold it to. Support also cannot repair your local path, whether that is congested Wi-Fi, a router placed badly or a loaded evening on your own line, and it cannot make an old set decode HEVC. Two settings do help and are worth trying first: buffer set to none, and a public DNS resolver.

Test that during the trial, deliberately. Send one question you already know the answer to and time the reply, then ask something specific, such as how many simultaneous connections the plan allows or exactly what the refund window excludes. Written answers matter more than fast ones, because a written answer is something you can quote later. Watch the shape of the reply too, since a script that ignores your four facts predicts how the next outage will go. Activation here takes about five minutes, plans refund inside 7 days, nothing auto-renews and no card is stored, which removes several conversations before they start.