Github user nickwallen commented on the issue:
https://github.com/apache/metron/pull/834
I clearly see the advantages of failing fast here, but I do have a
hesitation. I haven't even totally convinced myself of this. Maybe you guys
can help me think down this line of thought and make sure it is a non-issue.
This seems like we are introducing a hard dependency between Indexing and
Elasticsearch. If Elasticsearch is not running, then I cannot start the
Indexing topology. What if I am not using ES, but just want to index into HDFS
(or Solr, ha)? Or what if I am need to customize my templates? Does this
introduce any pain points that we are not anticipating?
---
