Github user HeartSaVioR commented on a diff in the pull request:
https://github.com/apache/storm/pull/2241#discussion_r129727744
--- Diff: conf/defaults.yaml ---
@@ -231,16 +228,13 @@ topology.multilang.serializer:
"org.apache.storm.multilang.JsonSerializer"
topology.shellbolt.max.pending: 100
topology.skip.missing.kryo.registrations: false
topology.max.task.parallelism: null
-topology.max.spout.pending: null
+topology.max.spout.pending: null # TODO: We dont need this any more
topology.state.synchronization.timeout.secs: 60
-topology.stats.sample.rate: 0.05
+topology.stats.sample.rate: 0.001
--- End diff --
Likewise I said some frameworks even don't sample and measure all for
meters. JStorm is one of example, and they (Alibaba) said "every day the JStorm
Cluster handles 1.5 PB/2 Trillion messages" in 2015, which is 23,000,000
messages per seconds. Yes we don't know how many servers they use, but they
also claim their performance in website http://jstorm.io:8080/Performance/ and
metrics seems not make considerable bottleneck.
One thing to consider is that current metrics implementation may be really
slower than dropwizard thing, so we may want to adjust rate if we believe it is
really a bottleneck.
Btw, most of the cases users would want to have operational convenience
over performance. We should consider first what users are facing day by day.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
