Description of problem: Potential issue with throughput when having labs across the WAN all pushing logs back to a central scheduler. There should be knobs/controls on the test harness or lab controller to limit how much data it will accept to push up to the scheduler. <fantasy-not-sure-if-useful-idea> if it was useful then a test writer could indicate which part of the logs it was interested in. ala tail or head. </fantasy-not-sure-if-useful-idea>
At the moment Harness is serializing all requests to LC, thus if LC has high latency, it will slow down. If this is not enough for communication between LM and LC, we should use traffic shaping on LC. Test writer can indicate which part of log is interesting: just has to use head or tail in his (or her) the tests.
Bulk reassignment of issues as Bill has moved to another team.
Dan, can we close this one based on the HTTP upload size limiting going into 0.17.1? The RFE also seems to predate logs remaining local to each lab.
Yes, the original concern about sending all logs from LC -> server over (potentially) slow WAN links is no longer relevant, we now store all logs within the lab until beaker-transfer comes along after the fact to move them.