I think we need more read+write mixed workload benchmarking to go on since these sometimes tease out locking problems.
-
-
Replying to @kellabyte
@kellabyte the lack of those benchmarks was what bit me when I tried to use RabbitMQ. “Step 1: Write, Step 2: Read” is not the same thing.1 reply 0 retweets 0 likes -
Replying to @kellabyte
@kellabyte the service would go unresponsive for minutes while it tried to GC itself under its memory limits. It was rough.1 reply 0 retweets 0 likes -
Replying to @kellabyte
@kellabyte …unless you turned off the producers. (which were people sending us things over the internet, so that’s hard to do)2 replies 0 retweets 0 likes -
Replying to @kellabyte
@kellabyte once it got into a bad place the consumers couldn’t ever keep up because rabbitmq wouldn’t give them messages fast enough anymore1 reply 0 retweets 0 likes -
Replying to @kellabyte
@kellabyte eventually I found twitter’s kestrel project and that has worked great for the last couple years.4 replies 0 retweets 0 likes
@lindvall @kellabyte speaking from the experience of running kestrel for twitter, it's still good to set bounds for kestrel :)
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.