Hey <!everyone>, Elementary CLI v0.18.2 Elementary CLI v0.18.3 is live! 🚀 Here’s what’s new: 🔔 Grouped Alert Subscribers Add subscribers to grouped alerts for streamlined, customizable notifications. 🔌 ClickHouse Integration Requires dbt package v0.18.3. 🌐 Universal Report Compatibility Reports now work across all languages and browsers — no more surprises. 🐞 Fixes & Enhancements •

--s3-acl

option for S3 report permissions — thanks @Luis Ventura! • Internal dbt project setup fix — awesome @Abhipal Singh! • Option to disable logo print — nice one @Nick Zeolla! 📦 To upgrade the CLI:

pip install elementary-data --upgrade

Elementary dbt package v0.18.3 is out as well! 📁 ClickHouse Artifact Support Initial support is here! Anomaly detection is not yet supported. Requires CLI v0.18.2. 🔥 Fix: Databricks Temp Relations Resolved issues with temporary relations. Let us know if you have any questions or feedback. As always, thanks for being part of the Elementary community! 🧡

Hey <!everyone>, meet our new hires 👋 We’ve just welcomed a few new members to the Elementary team. They handle the tedious parts of data quality — from recommending missing tests to writing metadata and finding root cause. 🧠 Meet them here! (Yes, they’re AI Agents. No, they don’t take PTO) But we think the team might still be missing a few key roles… 💡 Who else would you add to the squad? Drop your best AI Agent idea in the thread: • Name • Specialty • (Bonus: their LinkedIn-style career highlight) 🎁 Best ideas will get some swag — and maybe even make it into the platform.

<!here> 🚀 Hey Elementary community! We’re thinking about building a Python SDK to expand beyond dbt, with the same ease of use and code-first principles you know from our dbt experience. python We’d love your input! Got a minute? Tell us what you need ➡️ https://tally.so/r/w8ykOx

Hey <!everyone> ! 👋 You might’ve seen the teaser a few weeks ago — and now it’s official: Ella, our AI Agents Layer for Data Observability, has launched! 🤖 Ella includes 5 agents to help with key parts of your workflow: ✅ Test Recommendations Agent – Recommends and writes tests based on your policy, coverage, and lineage 🧠 Governance Agent – Flags & fixes gaps in ownership, documentation, and tagging 🛠️ Triage & Resolution Agent – Pinpoints issues and opens a PR with the fix 📚 Catalog Agent – Makes metadata instantly searchable and useful 💸 Performance & Cost Agent – Spots performance issues and suggests what to rewrite or retire They’re contextual, interactive, and built to get real work done — not just make suggestions. You can learn all about them in our docs, or watch the video below! If you're curious about what we’re building and shipping, feel free to join our #C08UM2MHQ7N channel. No commitment — just an easy way to stay in the loop. Hope to see you there! ✨

📣 Live Demo: See Elementary’s AI Agents in Action Hey <!everyone>! We’ve been talking about AI agents for observability—now it’s time to show you how they actually work. Join us for a live demo of Ella, our new AI-powered layer for data quality and observability. While this isn’t an OSS feature, it’s a chance to see how Ella connects with your Elementary setup. We’ll walk through real workflows and show how the agents help with: ✅ Recommending and writing tests ✅ Investigating data issues and surfacing root cause ✅ Improving documentation, tagging, and ownership ✅ Optimizing performance and cost ✅ Making metadata searchable and usable 📅 Wednesday, June 25th, 11:00am EST 🔗 Sign-up here: https://www.elementary-data.com/webinar/meet-ella-ai-agents-for-automated-data-observability See you there! 🔥

Hello <!everyone>, we just published a new behind-the-scenes post from our team! 🛠️ Streaming chat with AI sounds simple—until you try to build it. Most frameworks don’t natively support frontend-backend communication for real-time chat. So our very own @Ofek Weiss built a Python-native backend to power Ella, our AI agent layer for data quality—and broke down the full architecture here 🔥 And, we’ve open sourced it → 📦 https://github.com/elementary-data/py-ai-datastream (Feel free to ⭐_️ the repo if you find it useful!)_ Got questions? Ofek will be happy to answer them right here in the thread! 👇

<!here> 🚨 The AI Agents Live Demo is tomorrow, and the registration has been overwhelming. We’re showing how data teams are: ⚡️ Resolving incidents in minutes, not hours 🧠 Surfacing test coverage gaps automatically 🏷️ Closing metadata gaps with zero manual work 📅 Join us live → https://www.elementary-data.com/webinar/meet-ella-ai-agents-for-automated-data-observability (We’ll share the recording only with registrants) See you there! 🙌

Hey <!everyone>, Elementary dbt 0.19.1 + CLI 0.19.3 are here!🚀 Here’s what’s new in this release: 🔍 Zoom in on anomalies — improved anomaly drill-down experience ✨ dbt groups are now saved as artifacts - better visibility and structure for group-based workflows 🔔

edr monitor

now supports an

--exclude

option - skip models or sources more easily 📈 New

seasonality

parameter in

test_event_freshness_anomalies

- more flexible detection for seasonal patterns 🧰 Improvements & Fixes: • Added counters to table filter dropdowns • Fixed consistency bug with tags and owners You can upgrade with:

pip install --upgrade elementary-data

And don’t forget to update the dbt package in your

packages.yml

, then run:

dbt deps
dbt run --select elementary

🧡 dbt Fusion will be supported soon — we’ll keep you posted! As always, feel free to reach out with questions or feedback 💬

🚀 Update on dbt Fusion Compatibility Hey <!everyone>! Quick update on dbt Fusion: the Elementary team is prioritizing support for the integration and has made the necessary changes on our side. dbt Fusion is still in beta, and dbt Labs is actively working on adding support for their top packages—including Elementary—and we’re in close communication as they move things forward. We’re hopeful things will be resolved in the coming weeks and will keep you posted as progress continues 🧡 Thanks for your patience! 🙌

<!everyone> 💫 We’re excited to introduce the Elementary MCP Server — a new way to connect your Elementary context to any MCP-enabled tools like Cursor, Claude, or internal agents. It acts as a bridge between your live environment (models, tests, lineage, incidents, data health and test coverage) and the tools where work actually happens. Whether you're reviewing a data pipeline in Cursor, checking a dashboard in BI, or exploring data assets — Elementary can provide the full observability context and help you take action, all in natural language. Some examples of what it enables: • Prevent downstream breakages by checking column-level-lineage impact from source to BI during development • Get effective test recommendations based on actual usage and coverage gaps • Investigate issues with agents or copilots that have the full lineage and asset context • Understand any asset's health and current issue status without switching tools It’s not just about AI — it’s about better workflows, fewer surprises, and proactive reliability. ✨ Learn more about the our MCP server ✨ Sign up for the live demo to see it in action Available now on Elementary Cloud. We invite you to join the live demo and see what it can unlock for your team.

<!here> Tomorrow’s the day 🙌✨ We’re doing a live demo of the Elementary MCP Server, bringing data reliability and observability right into tools like Cursor, Claude, or even your own agents. We're getting great feedback from our Beta users do far: “I used the MCP to instantly see which alerts were the noisiest and check which critical assets would break if I changed a model — all without leaving Cursor.” “We integrated with GitHub and MCP — the setup was super slick. Excited to see Claude writing code with Elementary’s remote MCP.” 👀 If you want to see it in action → join us live: https://www.elementary-data.com/webinar/live-demo-elementarys-mcp-server

🚀 Hey <!here>! We’ve got another webinar coming up—this one’s for all Elementary OSS users: Scaling with Elementary OSS. As you know, Elementary OSS brings powerful observability to your dbt projects—but as your data environment grows, new challenges appear. 👉 In this session, @Joost Boonzajer Flaes and I will share what we’ve learned from hundreds of OSS users: how to overcome common blockers and what it takes to stay effective at scale. You’ll leave with practical tips you can put into action right away. 📅 Tuesday, September 30 at 11am EST 🔗 Save your spot here! Join us live and bring your questions!

<!everyone> Make the most out of Elementary OSS as you scale 🌎✨ We’re looking forward to seeing you in our upcoming webinar on Tuesday, September 30th, 11am EST : Best Practices for Scaling Elementary OSS. We’ll cover the challenges that come up as teams, pipelines, and organizations grow — including test coverage, alert fatigue and collaboration across functions. We’ll share best practices and solutions you can put to use right away. 👉 Register here Can’t wait to see you there! 🎉

<!here> We’re heading to dbt Coalesce next week! 🎉🔥 You’ll find us at Booth #428, right in front of the dbt swag store. If you’re around, come say hi — we’ve saved some merch especially for the community 💜 Also… keep an eye out for some big OSS news coming very soon 👀

Hey <!everyone>! Big news – Elementary dbt package 0.20.1 and CLI 0.20.0 are out, and dbt Fusion is supported 🐕🔥 What’s new: 🧩 dbt Fusion support (beta) Fusion is a complete rewrite of the dbt engine, bringing faster performance, static analysis, and improved reliability. Since Fusion is still in beta, so is our support for it — and we’re working closely with the dbt team to close the remaining gaps. 👉 See all details and migration notes here: https://docs.elementary-data.com/oss/integrations/dbt-fusion ⚠️ Some syntax changes are required to make sure everything works smoothly — check the docs carefully before upgrading. 📦 dbt package fixes & improvements • Fix datetime casting macro for ClickHouse (thanks StevenReitsma! 🙌) • Fix Athena

DROP TABLE

parsing error 💬 CLI enhancements • Improved Teams alert formatting (thanks @Frank Tubbing [Beequip]! 🎉) • Removed usage of deprecated

-m

flag in dbt • Removed Databricks version support restrictions • Fixed PostHog tracking issue • A few more small fixes and report enhancements! As always, we’d love your feedback — try it out, share what you think, and let us know if you hit any bumps!

<!everyone> We’re excited to announce Elementary 2.0: The Data & AI Control Plane. The Data & AI control plane brings observability, quality, governance, and discovery into one platform. Built on a shared context layer and a set of agents that automate reliability work and enable business users to self-serve, while engineers stay in control of what gets approved. The goal is simple: Give every team a consistent, tested, reliable foundation to work from. ✨ Learn more in the blog post: https://www.elementary-data.com/post/elementary-2-0-trusted-data-for-the-ai-era

<!everyone> Elementary dbt 0.21.0 and CLI 0.20.1 are live! kirby dance 🧡 dbt 0.21.0 highlights • New option to exclude detection period from training in all anomaly detection tests✨ • Fix for Redshift on-run-end failures • Error when columns cannot be detected (thanks @Denis Kokorin!) • Support for ClickHouse cluster mode (thanks Steven Reitsma!) •

failures

and

affected_rows

updated to bigint in the Elementary tables • Updated

require-dbt-version

for smoother compatibility *(*thanks @Jsolo!) magic CLI 0.20.1 highlights • Correctly shorten versioned model names (thanks @Ryosuke839!) • Fix for partial success handling introduced in dbt 1.9 microbatching (thanks @Nicola Corda!) • Fix BigQuery test result row scanning (thanks @Thomas Langton!) • Proper quoting of database and schema identifiers for SQL compatibility (thanks @mschmidoev!) The releases include many more small fixes and improvements. See the full release notes (dbt package, CLI) and upgrade to the latest versions! stocked New, helpful guides are now available: • Controlling on-run-end time and cadence • Full documentation for all Elementary schema tables Thank you to all contributors and users for helping us keep Elementary reliable, fast, and improving with every release.

Hey <!everyone>, Maayan here 👋🏻 I don't usually write in this channel, but this feels like the right moment. We started Elementary OSS because data reliability should be a standard, open and accessible. As teams adopted it and needs grew, we built Cloud to extend those foundations with automations, richer context, and features that connect engineering and business teams. Elementary 2.0 is the next step in that evolution. It brings together everything we learned from OSS and Cloud and takes it further: magic An OSS Python SDK to monitor any pipeline magic AI agents that remove repetitive reliability work from engineering teams magic Views and workflows that let business users self serve and contribute to reliability magic Multi project lineage for complex dbt ecosystems I would love to have you join me and @Or Avidov as we walk through where Elementary is heading. 📆 Jan 7th, 11 AM EST. Sign up here: https://www.elementary-data.com/webinar/introducing-elementary-2-0 Hope to see you there, and until then, happy holidays 💗🎄

It was the night before Christmas, and all through the stack, Not a pipeline was breaking—no need to Slack. The tests were all passing with care, In hopes that Q1 data would soon be there. Happy holidays <!everyone> from Elementary! Here's to ending the year with green tests and zero incidents. See you in 2026 when we'll help make sure your data doesn't have a New Year's hangover 🎊

Hey <!everyone>! ☃️✨ We love seeing Elementary show up in your LinkedIn job descriptions — So Elementary is now a LinkedIn skill stocked Add it to your current or previous roles and let the world know you’ve worked on delivering reliable, trustworthy data.

👉 Head to Elementary → Home → Products → Add as skill!

<!everyone> Elementary dbt package and CLI 0.22.0 are out! 🔥✨ This release includes a solid set of fixes, improvements, and community contributions. Huge thanks to everyone who helped make it happen 🙌 dbt dbt package 0.22.0 highlights • Fixed Redshift temp tables issue • Improved ownership inheritance from models to tests • ClickHouse cluster support (thanks @Steven Reitsma!) • Better support for hour-based detection periods • Enable

test_sample_row_count

config on tests (thanks @Ryosuke839!) 🚀 CLI 0.22.0 highlights • Support for Clickhouse alerts! • Add orchestrator jobs data to alerts (thanks @Lukas Steger!). This one’s exciting 🐕 • Python 3.13 support and replacing numpy with math (thanks Vugar Yagubov for both!) •

quiet-logs

flag (thanks @Mike Chrzanowski!) • Default

Config()

initialization now respects

target_path

in the default config file (thanks @ryanquincypaul!) 👉 Check out the full release notes and upgrade to the latest versions. ⭐ Love Elementary? A star on the repo is a great way to support the project and the community behind it! Thank you to all contributors and users for continuing to make Elementary better with every release.

Hi <!everyone> 🔔 Last call for the webinar today! 🔔 I'll be hosting the CEO of Flexor, they are turning unstructured data into context for AI. We will cover lessons we learned from working with teams that deploy analytics AI agent in production. Starting in half an hour, Join us 👇 https://www.elementary-data.com/webinar/how-to-prepare-your-data-platform-for-analytics-agents

Hey <!everyone>! Big release today. The demand was high, so we did an integrations blitz - and Elementary now officially supports 8 more adapters stocked 👉 New adapters: Microsoft Fabric, Vertica, DuckDB, and SQL Server. 👉Now fully supported: Dremio, Spark, ClickHouse, and Trino. All of this is available in dbt package and CLI 0.23.0, and on Elementary Cloud. If you've been waiting to use Elementary with your stack, or know someone who has, now's the time. 🚀 🔥 Book time with the team 📖 Quickstart OSS

Hey <!everyone> 👋 Many of you have been with us since the early days, when Elementary was just a dbt package. We started it believing data teams shouldn't have to work so hard for reliable data management. Last week we shared how that dream is becoming a reality: Our vision of the Agentic Data Team. AI agents that build pipelines, resolve incidents, safeguard data quality, and optimize costs. Many of these use cases are already in production with customers! We keep increasing the tools, capabilities and context for the agents. Our last product release included: • Code Review agent - Catches issues before they hit production, right in the PR. • Agents communication - Elementary agents can now send messages on Slack / Teams / Email, and open tickets on Jira / Linear. • Plan Mode - Agents show their work before acting, so you stay in control at scale. • New integrations: Fivetran, Lightdash, Snowflake Semantic Layer - Impact and root cause analysis, and context from ingestion to consumption. This community has shaped so much of what we've built. Many exciting releases for the OSS are coming soon as. If you want to hear more, I'd love to chat.

<!everyone> We're at DGIQ + EDW in San Diego! 🌴 If you're around, come meet @Maayan Salom and the team at booth 13. We'll show what the future of automated data management looks like. See you there 🙌

<!everyone> Elementary dbt package 0.24.0 and CLI 0.24.1 are out! 🚀 A good mix of new test options, alert improvements, and warehouse fixes. Three first-time contributors in this one — welcome aboard 👋 🧪 dbt package 0.24.0 highlights • New

min_value

parameter for anomaly detection tests (thanks @Caterina Andreu — first contribution! 🌱) • Capture dbt test descriptions in metadata (thanks @jdglaser — first contribution! 🌱) • Capture compiled code for microbatch models • Handle dbt deferral when retrieving the Elementary relation • Better anomaly detection on large time buckets — fixed

exclude_detection_period_from_training

• Fixed BigQuery nanosecond timestamp cast failure • Explicit

sqlserver__

dispatches for T-SQL / Fabric 📊 CLI 0.24.0 + 0.24.1 highlights • Native Slack Table Block for test result samples — cleaner, more readable alerts • New

skip_test_result_rows

parameter in

get_test_results

• Alerts with an unknown status are now skipped instead of crashing the run • Custom env var paths for the internal dbt project (thanks @Darkandz — first contribution! 🌱) 🔗 Full release notes are on GitHub — upgrade when you get a chance. ⭐ If Elementary is a true part of your day to day, a star on the repo goes a long way for the project and the community. Thanks to everyone who opened a PR, filed an issue, or upgraded early. See you in the next one.

Hey <!everyone>! 👋 AI is changing how data teams work, and the fastest teams are already rebuilding their day around it. We've got two live sessions coming up on exactly that. stocked June 24 · Putting AI to work: a day in Underdog Fantasy's data team How Underdog completely reworked their day around AI: catching quality and governance issues before production, managing the semantic layer, and keeping context current, all in dbt, Slack, the IDE, and CI. Real workflows you can take back to your own stack. 👉 Save your seat magic July 22 · Live Demo: Autonomous Data Management The one a lot of you have been waiting for. A live look at Elementary's autonomous data management: AI agents that build, operate, and govern your data stack end to end while your team supervises. The also watch how AI and dashboards consume your data, and work on the feedback loop that sharpens the semantic layer where answers fall short. 👉 Save your seat Hope to see you there, and bring your questions!

<!everyone> 🔔 Last call: Don't miss the chance to learn from Underdog's data team Camden's data team is moving fast on AI, and today he's walking through how they actually do it, all inside the tools they already use. Whether you're starting to automate data reliability workflows, or already building and monitoring context for AI, you're gonna want to be there. 📅 Today · 11 AM EST 👉 Save your seat

<!everyone> 📦 Elementary dbt package + CLI — v0.26.0 dbt package ✨ New test

metric_stability

— flags when a metric changes after you measured it (late or corrected data). You set how old a time bucket must be to count as settled and how much change is okay. Works with dimensions, custom time buckets, and

where_expression

. → docs.elementary-data.com/data-tests/metric-stability 🔍 Five new

_with_context

tests — same tests, now returning the failing rows plus nearby columns so you skip the follow-up query. → docs.elementary-data.com/data-tests/with-context-tests 🧱 BigQuery anomaly tests now run on nested STRUCT fields — no flattening needed. → docs.elementary-data.com/data-tests/anomaly-detection-tests/column-anomalies ⚡ Faster, cheaper runs — append-only result tables (no full-history scans), faster Fusion artifact uploads, dimension-partitioned anomaly training (fewer misses and false positives), and time-bucket names in non-dimensional alerts. CLI 🚀 dbt 2.0 and Fusion support. 🔧 Fixes — large test-result samples truncated so alerts don't fail on oversized payloads, more accurate alert timestamps, report UI fixes, and dbt

REUSED

state support (State v1.11+). 🌱 First contributions from @tlangton3, @zerafachris and @hiromi-dev — thank you!