Head's Up! These forums are read-only. All users and content have migrated. Please join us at community.neo4j.com.
04-07-2019 11:18 AM
On the page “Neo4j and Elastic Search” at the bottom there’s a section about Logstash plugins written by Pere Urbon. But the section is empty bullet points. A brief look at Github I didn’t find them. Has anyone thought about using Neo4j as the datastore for the Elastic Stack? Seems like if you have a good semantic model for all the things ES logs, you could get a lot of insights into a cloud application.
04-07-2019 11:25 AM
Ok, here it is:
05-14-2019 02:23 PM
Well... I’m in JRuby configuration hell. The plugin uses a few different libraries than the latest packaged with Ubuntu 16.04. If anyone has a recommendation for that, let me know. Also, not sure if the plugin. This project seems to be really comprehensive (although it does not have logger integrations):
05-28-2019 10:23 PM
Interesting idea, I guess so far no one has brought this up. Most work has been in the opposite direction,
also more recent ones like this from @david.allen
05-29-2019 08:19 AM
When I found that article, I knew that I wasn't alone in the world 😉
Elastic seems to be a terrific solution -- I don't have experience using it in a production environment. But there are a lot of links, ultimately would like to be able to trace those back to commits and lines of source code, configurations, etc.
06-30-2020 01:31 AM
hey, I have just found this topic. We have a logstash for hosting data ingestion pipelines. From kafka, files, jdbc to other databases, mainly SQL servers. I was looking for a new ingestion where I would like to load the Neo4j somehow, so I am looking something like this Neo4j output plugin.
@hank do you have any experience with the plugin so far? I am not sure it is maintained properly.
07-08-2020 07:18 PM
Hi Janos,
My brief look at the plugin, and that was a year ago, is that it is limited and it was using an old version of the API. So I lost interest...
That said, I think Logstash has such a wide array of collectors -- if you could push events, then go back and create other relationships. I recall the pattern for time series data is simply one node connected to the next. So I never thought it could, in real time, create the kind of relationships that squash down all the events into something meaningful.
It did not look too difficult to write a plugin for TS data, but it was out of scope if there were not supported tools.
Let me know if you take it up.
Best,
Hank
i/o werx
All the sessions of the conference are now available online