I switched my setup today over from machines running the remote harvester to sharing disks over the network (posted about it in more detail here) and noticed in my logs I have some isolated cases of checks exceeding the 30 second limit:
Another bottleneck to consider may be not to harvest from disks that are currently being written to after plotting? I think that’s the most likely reason in my case, though I’m going to keep an eye on the timing of this. I’m not transferring any plots over the network in my case also so there shouldn’t be any issues in terms of it being saturated