Timescale

At Timescale, time-series data is treated as a first-class citizen — metrics, events, and system behavior are stored so teams can understand how things change over time. Because most outages don’t come from one bad deploy. They come from small degradations nobody noticed early. Good DevOps is less about reacting fast and more about seeing slow failures early. How do you track performance trends over time in your stack? #DevOps #ServerScribe #Timescale #Databases #Monitoring #Performance

The #Postgresql community has built one of the world’s most mature and well-engineered databases, and it deserves AI tooling that reflects that depth. We are asking the community to help us build it. Today we’re launching pg-aiguide, an open source project that teaches AI coding tools to write real, production-quality Postgres. Right now, most AI assistants still generate SQL with classic pitfalls: • money for prices • BRIN on random data • mixed SERIAL and UUID • timestamp without timezone • case-sensitive email uniqueness The SQL works, but it leads to bugs, migrations, and performance issues down the road. pg-aiguide fixes this by giving AI the Postgres-specific judgment it has been missing through: • AI-optimized Postgres skills • Version-aware semantic search • Extension ecosystem docs The goal is simple: AI that writes Postgres you actually want in production. We are expanding the skill library and building deeper guidance for indexing, full text search, and extensions. We would love help from the #postgres community. If you maintain an extension or have expertise to share, please contribute. We are especially looking to add PostGIS and #pgvector expertise, but any Postgres-related tools and knowledge are welcome. Repo: https://lnkd.in/eDCVwVBv Let’s teach AI to write Postgres like an expert, together.

☁️ Announcing Tiger Cloud on Azure While Azure Postgres has long embedded TimescaleDB Apache 2 Edition, this lacks advanced capabilities for query speed and cost savings like compression and hypercore columnar storage. Teams split workloads between Postgres and ADX. Dashboards slowed down at scale. Today that changes. Tiger Cloud on Azure brings the full TimescaleDB experience to Azure as a managed service: ✅ Hypercore columnar storage for fast analytics  ✅ Incremental continuous aggregates for real-time rollups ✅ Native compression that actually works  ✅ Hyperfunctions for advanced time-series operations  ✅ Stay 100% in PostgreSQL—no KQL, no pipelines Initial benchmarks between Tiger Cloud vs Azure Database for PostgreSQL show significant improvements for workloads mixing recent and historical data: 225× faster queries, 95% average compression, sub-10ms response times. Available now in: ▪️ East US 2 (Virginia) ▪️ West Europe (Amsterdam) More regions are coming based on demand. Ready to get started? Signup for a free trial at tigerdata.com. Enter your credit card directly (or get invoice billing if needed) via your Tiger Cloud console. Alternatively, if you want to pay through Azure Marketplace, select from two options: Pay-As-You-Go and Annual Commit. Your existing Azure setup stays the same. With Tiger Cloud on Azure, you no longer have to choose between fast analytics and simple architecture. Start your free trial: https://lnkd.in/gfthV3F6 #Azure #TimescaleDB #PostgreSQL #DataEngineering

Mechademy helps the world’s largest energy operators keep critical turbomachinery online, covering assets that include 6% of global LNG production. At that scale, every minute of downtime can cost millions. Mecademy’s hybrid digital twins help deliver consistent turbomachinery uptime by detecting early degradation, and prescribing fixes that improve uptime by 2–10% and deliver ~15× ROI across fleets exceeding 2.5 million horsepower of driver power. As digital-twin workloads expanded, Mechademy hit a performance wall: ▪️ Exploding compute on MongoDB: Even smaller tenants were running ~10,000 tests every 30 minutes with CPU pegged above 95% and query targeting >1,000 forcing frequent vertical scaling. ▪️ Operational drag: Nested aggregation pipelines and ad-hoc rollups turned every new diagnostic into a migration plan, not a configuration change. Mechademy needed a data layer built for industrial time-series at production scale without runaway costs. Tiger Data (creators of TimescaleDB) transformed their infrastructure: ▪️ Native time-series architecture: Hypertables replaced manual bucketing and schema wrangling; continuous aggregates delivered the right resolution for every test automatically. ▪️ Performance where it matters: On equivalent hardware, base-table queries ran 66% faster; continuous aggregates sped up by 18% (1-min), 81% (10-min), and 95% (1-hour), with far less data scanned. ▪️ Massive efficiency gains: Built-in compression slashed storage and boosted scans; Mechademy now processes 10 million diagnostic tests every 30 minutes on an M20-class TimescaleDB cluster. Built for scale and savings: ▪️ 87% reduction in infrastructure costs ▪️ 50× increase in workload capacity (200k → 10M tests per half hour) ▪️ Near-zero maintenance overhead with hypertables + compression The result: Mechademy’s hybrid digital twins run with industrial reliability and cloud efficiency, turning a maintenance burden into a measurable advantage. Read how they did it: https://lnkd.in/g_Q7PMza

Agents are the new developers. And they need databases built for how they work. Today we're open-sourcing Eon: a production-ready Slack agent that our team actually uses regularly. After 6 weeks, nearly 50% of the company depends on it for instant answers from our institutional knowledge base we’ve generated. Building Eon taught us what agents truly need to operate in the real world. We faced three fundamental challenges: ➡️ Conversational memory – Agents need to follow threaded context, not just respond to isolated messages. We built real-time Slack ingestion into TimescaleDB, treating conversations as time-series data. ➡️ Focused context, not generic tools – Official MCP servers expose dozens of tools across every API endpoint. That flexibility creates problems: wasted tokens, cognitive overload for the model, and more errors. We built focused servers with just the tools users actually need.. ➡️ Production reliability – When half your company depends on an agent, "mostly working" isn't enough. We built durable event processing with automatic retries, bounded concurrency, and millisecond latency. We're releasing everything as open-source, composable components: ▪️ tiger-eon – Our reference implementation, lightly edited from production ▪️ tiger-agents-for-work – Production-ready framework with durable queues and retries ▪️ tiger-slack – Real-time conversational memory in TimescaleDB ▪️ tiger-docs-mcp-server – Semantic search with auto-discovered prompt templates ▪️ tiger-gh-mcp-server – Focused GitHub tools for PRs and issues ▪️ tiger-linear-mcp-server – Streamlined Linear integration Each component works standalone or together. Deploy with our interactive setup script in ~10 minutes. Built on Postgres. Try it today: https://lnkd.in/gkqJ98v4

This week at Tiger Data (creators of TimescaleDB) we launched Agentic Postgres, the database for agents. Here’s what launched and where you can read all about it. In Postgres for Agents, Tiger Data co-founders shared their vision and an overview of what we’ve built: https://lnkd.in/gpbrGH6x Here’s where we introduced the Agentic Postgres Free Plan (the fastest way to experiment with AI on Postgres): https://lnkd.in/gSB9_Juw And since agents don’t just query data but also provision, fork, and learn, now they can do it all on Postgres through: ✓ MCP server + APIs + CLI for autonomous workflows: https://lnkd.in/gUbC9a2E ✓ Fast zero‑copy forks for safe, quick sandboxes: https://lnkd.in/gnyQcXr6 ✓ Hybrid BM25 + vector search: https://lnkd.in/gUCRpn8w For developers who like to learn by watching, here are the videos: ▶️ How to Safely Test AI Database Workflows With Tiger CLI + MCP: https://lnkd.in/g39Gjz4M ▶️ Introducing Prompt Templates - How We Taught AI to Build Better Databases: https://lnkd.in/gH5hpXcU ▶️ Zero-Copy Forks: There is Now a Safe Way to Test AI Migrations: https://lnkd.in/gsPbNfkZ Next week’s preview: ➡️ Fluid Storage - a deep dive into the new distributed storage layer powering Agentic Postgres  ➡️ Tiger Agent for Work - a production-ready library and CLI for building Slack-native agents for serious work.

Your coding agent just generated a database migration. Tests pass. It's Friday at 4pm. Do you deploy it? Most developers wait until Monday. The agent worked on test data—you don't know what happens against 40 million real rows. That's why we built forks into Agentic Postgres. Fast, zero-copy database forks let you test AI-generated migrations on real production data—safely isolated, spin up in minutes, pay only for what changes. → Catch edge cases before they break prod → Debug performance on actual table sizes → Deploy with confidence (yes, on Fridays) Read the deepdive we just put out! https://lnkd.in/gTF-TuGi

AI is changing how we write code — faster, riskier, more ephemeral. But our databases haven’t kept up. Your AI agent can generate migrations, optimize queries, or rewrite schemas — yet your “test database” has 100 rows, and staging is weeks out of date. That’s why we built fast, zero-copy database forks for Postgres. Fork your production database in minutes. → Test AI-generated migrations on real data → Debug performance issues safely → Pay only for what you write Built on copy-on-write storage, forks share underlying data with the original — no full clone, no extra bills. It’s how developers move faster with real data and deploy with confidence (yes, even on Fridays). Available now via Tiger CLI and GitHub Actions. Read the full deep dive → https://lnkd.in/gwJBWHD2 (And see below the post about it for more perspective by Tiger Data (creators of TimescaleDB) Head of Developer Advocacy and Docs Matty Stratton.

𝗔𝗴𝗲𝗻𝘁𝘀 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗡𝗲𝘄 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿 Agents, like Claude Code, feel uncanny. My first time using it, I built a mobile web app that tracked pushups using computer vision. Just for fun, to see what it could do. One hour later (mostly its time, not mine), I had an app that just worked. That gave me goosebumps. For the first time, it felt like software wasn't something I built, it was something building with me. It felt like something brand new. I realized: agents had become the new developer. But software agents don't behave like human developers. Software development tools need to evolve. Agents need a new kind of database made for how they work. So we built it. And we're launching it today. 𝗔𝗻𝗻𝗼𝘂𝗻𝗰𝗶𝗻𝗴 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗣𝗼𝘀𝘁𝗴𝗿𝗲𝘀: 𝗧𝗵𝗲 𝗙𝗶𝗿𝘀𝘁 𝗗𝗮𝘁𝗮𝗯𝗮𝘀𝗲 𝗕𝘂𝗶𝗹𝘁 𝗙𝗼𝗿 𝗔𝗴𝗲𝗻𝘁𝘀. There's a lot of engineering behind this: a new copy-on-write block storage layer, fast zero-copy forks, new Postgres extensions for full text search (BM25) and semantic search, what (we think is) the best MCP server for Postgres ever built, and a new CLI and free tier. I'm very proud of what this team has built, especially in such a short period of time. More here: https://lnkd.in/eWHY8Quk Please give it a try. We're just getting started. We'd love your feedback. 🐯 🚀 cc Tiger Data (creators of TimescaleDB)