Elastic{ON} Tour Frankfurt 2018

To be honest, I felt a bit sad when Elastic announced the end of the world user conference in San Francisco. I was last year in San Francisco and it was really a great experience. But of course, I understand that it not possible for everyone to join a conference at the US west coast. The concept of the Elastic tour brings the event to the people instead of inviting the people to the event. I like that idea.

The one-day conference was an interesting mix of announcements, deep dives in the elastic stack and use case presentations. In addition to that, there was an ask me anything (AMA) booth where I had a really nice discussion with some Elastic employees about multi data center setups. Last but not least, there was a a demo station where you could get in touch with almost every feature of the whole product stack.

I really enjoyed the day in Frankfurt. The talks that I attended were interesting and had the right mix of business and technical details. My personal highlights:

The Elastic Common Schema defines a common set of attributes for ingesting data into Elasticsearch. It helps to correlate data from different sources and aims to reduce confusion because of naming issues. As history told us: conventions and standards are helpful. [1]

Adaptive replica selection is part of Elasticsearch since version 6.1, but deactivated by default. Generally, Elasticsearch selects the nodes that will be queried during a search request in a round-robin approach. That might involve often nodes that respond slower than others. To avoid slow queries, it would be helpful to select nodes in a smart way, that prefers fast responding ones and that is what adaptive replica selection does. There is a good blog post from Elastic that explains it in detail. [2].

Frozen indices introduce a new option to handle old, less frequently used data. It has a good ratio from disk space to heap usage with the cost of a higher search latency. It might become an important part of a index management strategy in addition hot and warm indices. [3]

Kibana creator Rashid Kahn demonstrated the current state of Canvas. It was nice to see how it evolved since I saw it for the first time at Elastic{ON} 2017. Imagine Canvas as a kind of web-based presentation tool. But with interactive content powered by Elasticsearch data. I really hope that it will be released soon. It looks very promising. [4]

And last, but not least all the interesting conversations with other visitors and Elastic staff. If you ever have the chance to join an Elastic event, don’t hesitate to do it. It’s worth it.

Add a Comment

Your email address will not be published. Required fields are marked *

By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.