Tiger Data (creators of TimescaleDB)

Tiger Lake is now in public beta for scale and enterprise users. Finally, a real data loop between Postgres and your lakehouse. Tiger Lake is a native Postgres-lakehouse bridge for real-time, analytical, and agentic systems. No more stitching together Kafka, Flink, and custom glue code. Tiger Lake creates continuous sync between Postgres and Apache Iceberg on S3, built directly into Tiger Cloud. It streams any Postgres table to Iceberg via CDC, and can replicate existing large tables from Postgres to Iceberg via optimized backfill transfers. No need to choose between operational speed and analytical depth. With Tiger Lake, you get both in one architecture. Details: https://lnkd.in/e98mbXfK

Last month, we shipped our biggest AI launch yet, which is the new MCP server in Tiger CLI. With such a big product launch, it's easy to miss things. So I went ahead and shared with you five features you're probably not using with the new Tiger MCP, but really should! 1. Let your AI manage your databases Run `tiger mcp install` and your AI assistant can create services, check connections, and run queries without you leaving your editor. 2. Instant database forks Zero-copy clones in seconds. You can fearlessly and easily test migrations against real data, then delete when you're done. The first anxiety-free Postgres platform for agents. 3. Docs search built-in Your coding agent can search Postgres docs (versions 14-18) and TimescaleDB docs directly. No more browser tabs. Always up-to-date. 4. Skills that teach your AI Postgres best practices Alongside Claude's launch of Skills, we also created our own after distiling 30 years of Postgres know-how. Schema design, hypertable setup, compression policies. The stuff that prevents painful migrations later. 5. Run SQL through your AI Ask "how many events came in today?" and get the answer. No need to remember SQL connection strings. All within your coding agent. If you're using Tiger Cloud, all of this is already available: tiger mcp install Pick your editor (Claude Code, Cursor, VS Code) and you're set. Full blog: https://lnkd.in/e4tbxEHF

✨TimescaleDB 2.24 feature teardown: Lightning-fast columnar compaction TimescaleDB brings an optimized columnstore to Postgres, but one that still looks like a standard table. Under the hood, a hypertable is partitioned into chunks: the most recent data typically in row format, the rest in compressed columnar form. Query planning hides all of this, however, and users simply see a table. Unlike many columnstores, ours is fully (and transactionally) mutable. Inserts, updates, upserts, and deletes all work efficiently. Each columnstore chunk is paired with a tiny interim rowstore: new mutations land there, and queries read from both transparently. Over time, these changes are written back to the columnstore as larger, well-ordered batches. And in high-ingest workloads, Direct Compress writes compressed batches directly to the columnstore, bypassing these interim rowstores entirely. Both write paths can produce many small columnar batches, especially when a segmentby key creates separate batches per tenant, device, or id. These batches are correct, but suboptimal for locality and scan efficiency. Compaction merges and compacts these fragmented batches into larger, better-organized segments. In 2.24, this now happens fully in-memory (respecting maintenance_work_mem), avoiding the older disk-based sorting/segmentation path. The result: 4-5x faster compaction far lower I/O, cleaner physical layout, better locality, and faster queries. All while behaving like a single Postgres table. 🐯🚀 Tiger Data (creators of TimescaleDB)

✨ TimescaleDB 2.24 feature teardown: Direct Compress just got smarter (and faster) When we introduced Direct Compress in 2.21, the idea was straightforward: instead of writing data into the row store and later converting it to compressed columnar format, perform that transformation on the ingestion path. For high-ingest workloads, we compute a compressed batch at write time and persist it directly into the column store as a single transaction. In 2.24, we’ve extended this to work seamlessly with hypertables that maintain continuous aggregates. TimescaleDB has always collected invalidation ranges in memory during a transaction and flushed them at commit. These ranges identify which pre-computed aggregates must be refreshed in response to the new data. This release extends that same mechanism to Direct Compress, ensuring that invalidation ranges reflect the new compressed batches being built during ingestion. The result is that users can now combine two of TimescaleDB’s most important capabilities -- the hypercore columnar engine and continuous aggregates -- while retaining the benefits of Direct Compress: lower write amplification, meaningfully faster ingest under load, and smaller WAL footprints with fewer IOPS. We’re building the best #Postgres. Now even faster and easier. 🐯🚀 Tiger Data (creators of TimescaleDB)

🚀 That’s a wrap on AWS re:Invent! Our booth was buzzing with energy from start to finish. It was great catching up with customers, partners, and friends - old and new. Huge thanks to everyone who stopped by, joined our talks, and attended our happy hour! We loved connecting with you all and sharing how Tiger Data is the leading time-series Postgres solution for real-time, analytical, and agentic applications. See you at the next one! 👀

Why can't agents find the right docs, and how does Postgres fix it? This was the topic of the latest talk by Tiger Data (creators of TimescaleDB)'s own DevRel 🤖 Jacky Liang at AI Dev 25 x NYC, where he showcases why AI agents struggle to find the correct documentation and how to fix this using PostgreSQL: https://lnkd.in/giXDTbWc

How Flogistix by Flowco cut infra management costs by 66% with Tiger Data Flogistix by Flowco uses streaming data to optimize vapor recovery and maximize well performance. But their tech stack couldn’t keep up. They were stuck with: ▪️ Fractured mix of technologies ▪️ Limited expertise on fine-tuning performance ▪️ Lagging regulatory follow-ups  ▪️ Delayed field service insights ▪️ Exponential growth in data They needed a system that: ▪️ Their team could run with existing Postgres/SQL skills ▪️ Scaled cleanly for time-series data ▪️ Exceeded customer SLAs ▪️ Gave field techs real-time insights on-demand Enter Tiger Data, Flogistix eliminated their worst pain points around cost and query latency, and saw:   ▪️ 66% cost savings on infrastructure and storage ▪️ 84% data compression in production ▪️ Uptime jumping from ~95% to “Well above 99%” If you’re running data-heavy industrial workloads and feel your stack straining…this case study is worth a read. 👇 🔗 https://lnkd.in/gmHPmDKb #TigerData #TimescaleDB #Flowco #Flogistix #NoSQL