On Sat, Feb 14, 2015 at 6:56 PM, Tim Bain <[email protected]> wrote:
> Are you really looking for every message to result in a second message > going to the advisory topic, especially at the volume you've described for > your use cases? This seems like a good opportunity to rethink this part of > your solution (and we'll help of you'd like). > I appreciate it. At the moment I’m fine in terms of raw message throughput. My issue so far has been the performance with LevelDB since I’m using somewhat ephemeral queues. It takes about 100ms to GC them which is my big problem at the moment. I switched to using an in-memory broker and so far my performance has been fine. This is why I asked about the design of LevelDB there and see if there are any obvious fixes. Kevin -- Founder/CEO Spinn3r.com Location: *San Francisco, CA* blog: http://burtonator.wordpress.com … or check out my Google+ profile <https://plus.google.com/102718274791889610666/posts> <http://spinn3r.com>
