I want to locate the delay before changing the network. Our read-only FR10 display freezes when detailed logging writes to the shared folder, then catches up. Cell controls are separate. The programmer suspects file writes, but we have no timing around the request itself.
Time the request, the processing and the file write separately, with the timing capture kept away from the same slow writer, so the measurement does not disappear into the delay you are investigating.
Programmer added separate timings. In the captured freeze, the request returned before a long file write. We will compare more than this one occurrence before changing the display.
Have the programmer review a bounded logging queue and an explicit policy for a full queue or failed writer, because moving the writes elsewhere without a limit can simply move the eventual stall.
Anil, yes, it did. The proposed display shows reading age and logging health separately. Sarah, the queue limit and overflow behaviour are in review; the revised writer has not been installed yet.