build an intelligent query layer with MCP in #elasticsearch โ from search to agent
here's a preview how ES|QL queries turn into tools that get aggregated into agents โ all served and built right in kibana. and you can try this out today
We went for an AGPL license because we didn't want the elastic problem.
While JP Morgan can run @daytonaio internally for free, Amazon can't spin it up and run a competing service without open-sourcing everything they build on top.
The ideal long-term goal is to separate the truly permissive core (MIT or Apache) from the enterprise-scale features. Individual developer value in one repo, and compliance + audit logs + reporting in another.
You can be genuinely open source in the parts that matter to the community and protect the parts that only matter at scale.
We didn't have time to do that cleanly on v2, so AGPL was the fastest defensible choice.
@zbuckholz@ProtonMail good old tribe node. legend has it that this feature was written on one transatlantic flight (and some people would then say "it showed" ๐ )
but cross cluster search should be a lot more robust nowadays
@quienesmhulhar that's one way of solving it (and not worst :D)
but if you run multiple processes โ especially memory hungry ones โ you might still want to look at the JVM heap configuration
@quienesmhulhar but are the processes all running on the same instance? you'll need adjust the memory settings for elasticsearch (and probably logstash) in that case
@srchvrs@elastic I think primarily hierarchical k-means (instead of k-means) and SOAR. https://t.co/xmZGNZ46KJ has a deeper dive into the ideas behind it
plus better (binary or more) quantization: https://t.co/lhLc1TXzer
Super excited to be live demoing @asymptotelabs' Agent Beacon at @elastic's Developer Meetup on May 28 in San Francisco.
Iโll walk through how Agent Beacon turns AI agent runtime activity into Elastic-native security investigations, helping teams connect suspicious behavior back to the agent workflow behind it.
Iโll be presenting alongside lead engineers from @recurly and @Intuit for a night of real production stories across search, reliability, observability, and AI runtime security.
If you're around, please swing by! Luma link below in the comments ๐
We are investigating unauthorized access to GitHubโs internal repositories. While we currently have no evidence of impact to customer information stored outside of GitHubโs internal repositories (such as our customersโ enterprises, organizations, and repositories), we are closely monitoring our infrastructure for follow-on activity.
@planedrop I can't speak for the elastiflow docs but for a simple elasticsearch setup I'd look at https://t.co/TeIE4IuPhS โ it's one curl command to get everything up and running on docker (or podman)