Frame-loss lines in the log are a recovery counter

Synchronized data travels as a chain: every update the server sends names the update it sent before it, so a viewer that misses one notices immediately and asks for a fresh copy. That recovery has been in place for a while, and it works — but it writes a line to the log each time, and a large number of those lines has repeatedly been read as a large amount of lost data.

It is the opposite. Each line is a gap that was detected and already repaired. The data-sync architecture page now says so, and gives the two numbers that actually diagnose something: how often a single stream logs it, and whether a fresh copy ever follows. It also lists the log lines to read alongside it, because the cause is almost always something further upstream ending and re-establishing a subscription — not the chain itself.