Github user nickwallen commented on a diff in the pull request:
https://github.com/apache/metron/pull/984#discussion_r180154515
--- Diff:
metron-deployment/packaging/ambari/metron-mpack/src/main/resources/common-services/METRON/CURRENT/configuration/metron-enrichment-env.xml
---
@@ -129,70 +145,156 @@
<property>
<name>enrichment_topology_max_spout_pending</name>
<description>Spout Max Pending Tuples for the Enrichment
Topology</description>
- <value/>
- <display-name>Enrichment Max Pending</display-name>
+ <value>500</value>
--- End diff --
This is about tuning the user's day 1 experience. Specifically what I have
seen is on day 1 users push some telemetry into a topic, start up the topology,
and the topology just blows-up. They have no idea why because they are on day
1. That is a bad experience.
The topology blows-up because on day 1 the user is running with a topology
that has NOT been tuned. The topology blows-up because when this value is
undefined there is no back pressure. We need some back pressure here to avoid
the day 1 blow-up.
I'd prefer a day 1 user be able to start up the topology and see it
actually work, albeit slowly, under most conditions. A user running against
heavier load is more likely to be experienced and should have more knowledge to
know that this value needs increased.
---
