We are pleased to announce the release of Elastic 8.0.0-beta 1. This is part of a series of pre-8.0.0 releases designed to let you test the features and changes coming in 8.0.0 with your application.
The next chapter in search begins
With Elastic 8.0, we are excited to bring you the most performant, secure, and scalable Elastic Stack to date. These new enhancements not only arm you to solve your most challenging problems, but also amplify our turnkey Elastic Enterprise Search, Observability, and Security solutions, which are all built into one stack, the Elastic Stack. With the release of 8.0.0-beta 1, we want to give you a sneak peek at some of the highlights, including simple security, PyTorch NLP models now supported in Elastic machine learning, and our 8.0 upgrade assistant.
Simple security — on by defaultProper security has never been more important than it is today. And while security features have been free and included in Elasticsearch and the Elastic Stack for many years, we spent the last year working to simplify the steps to enable security so that no clusters should be run without it. In Elastic 8.0, security is on by default, and we have streamlined this process to save you time and ensure that your data, network, and users are secured in the Elastic Stack.
In 8.0, security setup is a fast and easy process with features like in-product assistance and auto-generated tokens and certificates that help to expedite configuration.
And we’re not stopping there. In the coming releases, we will continue to provide enhancements to stack security to ensure your data is protected while making security configuration easier to save you time. Interested in learning more? Watch our ElasticON session for further information.
Third-party NLP models now supported in Elastic machine learningThere are a few features — searchable snapshots and Kibana Lens — that fundamentally changed the way people use Elastic. The next paramount shift is support for third-party NLP (BERT-based) PyTorch models in the Elastic Stack. This new feature introduces a method of importing third-party models for native use within Elasticsearch, bringing modern natural language processing (NLP) capabilities to the Elastic Stack. NLP unlocks the ability to do name entity recognition (NER), semantic text embedding, emotion and sentiment analysis, or text classification and zero-shot classifications that can enhance your search use case.
Using Elastic’s Eland client library, you can upload your custom PyTorch models or use community models found on Huggingface.co. For 8.0 general availability (GA), we are including a compatibility matrix of supported models so you can find what will fit your search needs. And, if your data is not stored on Elasticsearch, we have integrations into other ecosystems like Hadoop with ES-Hadoop and Apache Spark.
Third-party PyTorch models will be in technical preview for Elastic 8.0.0 and will be GA in a future release. We would love your feedback — please join us on the Elastic Community Slack #machine-learning channel to talk about use cases or ask questions about this new capability or check out our ElasticON session on modern NLP.
Get ready for 8.0.0 GA with our Upgrade AssistantAs in previous releases, we’ve been hard at work developing an in-product tool to get you and your cluster ready to upgrade. The Upgrade Assistant, which will be available with our 7.16 release, shows which steps need to be completed prior to upgrading your cluster and identifies any deprecated settings in your 7.16 configuration. Some issues can easily be resolved with a click of a button; others may involve a multi-step process. But fret not — the Upgrade Assistant will guide you through the process of resolving issues and will also provide additional context to help you decide how best to move forward. If you would like to hear more, check out our ElasticON session.
8.0 is just the beginning
We’re excited to share the beginning of our journey into 8.0.0 and hope you will join us by downloading 8.0.0-beta 1.
We’d love to get your feedback and tips on any bugs you find through our Pioneer Program. To become an Elastic Pioneer, try out the preview of any (preferably, every) part of the Elastic Stack and open issues as you find them in the appropriate repo — Elasticsearch, Kibana, Beats, Logstash, and APM. When you open an issue, mention that you found the bug in 8.0.0-beta 1, and we’ll add a “Pioneer Program” label.
The release and timing of any features or functionality described in this post remain at Elastic's sole discretion. Any features or functionality not currently available may not be delivered on time or at all.