AWS Thinkbox Discussion Forums

beta 10 monitor performance

Thanks guys, its a visible improvement. We will let the update trickle out to production for a couple of days before asking them for feedback.

I am still getting almost constant 100% cpu usage. Does the monitor really need to use that much cpu / ram? It feels way too much. Some users started setting the cpu affinity on the monitor to only use the last core, because it affected their OS performance in other applications too. Or they keep monitor closed, and just open them when they need to check on their job status (and this is why they asked for a mini monitor, deadline 5 style. They feel the full monitor takes too much resources)

I think the problem is that the previous UI slowdown was introducing an artificial delay, so not as many updates could be processed. In addition, because it was in Python, it would only saturate a single core. Now that it’s all multithreaded, that could explain why it saturates all cores now.

We’re having some discussions internally about this today to try and figure out what we can do here. For example, we could do more synchronous updating each interval (ie: instead of having the jobs, slaves, etc update independently, they could be done in order at a set interval).

Privacy | Site terms | Cookie preferences