Our fixture-block telemetry display now stalls while the log folder is busy. This diagnostic application is separate from cell controls. Synchronous writes look suspicious, but I have not measured where the delay begins. How should I separate writing time from delayed incoming messages?
Can you time receipt, decoding, file writing and display separately, then replay the same captured input offline with logging on and off so the comparison isn't two different incoming streams?
The matched replay reproduces the pause while synchronous flush blocks the display thread; receipt timestamps remain regular in that test. We have isolated a local logging delay, not ruled out every live communication problem. The maintainer is reviewing a bounded writer queue with visible failures and shutdown handling before changing the deployed monitor.