[
https://issues.apache.org/jira/browse/STORM-3499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16917446#comment-16917446
]
Jay Bhavsar commented on STORM-3499:
------------------------------------
Hi,
My bolt uses ES bulk API in contrast to these two bolts which index each
document individually. Bulk indexing is significantly faster for large number
of records.
Your concern about storm's at least once processing is legit. However for my
particular use case, it was acceptable to trade some reliability for faster
processing. Maybe some others can relate to this.
> allow bulk indexing in elasticsearch
> ------------------------------------
>
> Key: STORM-3499
> URL: https://issues.apache.org/jira/browse/STORM-3499
> Project: Apache Storm
> Issue Type: Wish
> Components: storm-elasticsearch
> Reporter: Jay Bhavsar
> Priority: Trivial
> Labels: features, newbie
> Original Estimate: 504h
> Remaining Estimate: 504h
>
> For a particular use case, I created custom bolt that can push data to ES in
> bulk. It's quick and dirty code as of now. The bolt acks tuples as soon as it
> revives it, before sending to ES. (need more thoughts/suggestions on this
> behavior). Can we make this as feature of storm? I am willing to contribute.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
