Tiger Data (creators of TimescaleDB)

Postgres turns 40 this year. Despite having a name only a parent could love, it's still the database developers know and love most, even in the AI era. We asked five industry experts why. https://lnkd.in/gemY-Bvz #postgres Chris J. Preimesberger Stephen O'Grady Carl Olofson Paul Gillin

McKinsey found that 74% of manufacturers never get their IIoT pilots past the pilot stage. Upstream oil and gas looks the same. Everyone assumes it's a bad query, or bad hardware. It's usually neither. Nobody drew the boundary. Every database has a performance envelope: how much storage you can afford, how fast you can ingest, how slow your queries are allowed to get. Outside that envelope, things that worked fine at 50 wells fall apart at 500. Doug Pagnutti is doing a live session on how to calculate that envelope before you commit to an architecture, using a real well-monitoring workload (10,000 tags at 1Hz) as the example. We'll cost out a year of retention, find where the ingest wall actually sits, and look at what changes when you add TimescaleDB to Postgres instead of replacing it. Free, live, tomorrow at noon ET. Can't make it? Register anyway and we'll send you the recording. https://lnkd.in/e7hyxYW8

We're heading to IMTS 2026! If you're managing telemetry from CNC machines, sensors, or shop-floor systems, you already know the pattern: data piles up fast, and your database has to keep up without falling over or splitting into a second system. Come find us at booth #236355 in Chicago, September 14–19. We'll be talking through how Tiger Data keeps operational analytics live and correct on Postgres, no forced re-architecture required. See you at McCormick Place → https://lnkd.in/gjFsuGS6

Companion.energy's Azure PostgreSQL hit a wall: queries timed out and dashboards took 2-3 minutes to load. The team (Co-founder Jonas Verstraeten, Senior Software Engineer Tristan Vermeesch, and Infrastructure Engineer Gert-Jan Braeckevelt) needed skip-scan, compression, and continuous aggregates—features only Tiger Data could provide. After migrating to Tiger Cloud, the results spoke for themselves: - 25x faster queries (4,950ms → 202ms) - 98% compression (1.7TB → 39GB) - 5-minute database forks - Zero-downtime migration completed in 8 weeks One database now handles real-time ingestion, customer queries, and analytics. See how Companion.energy scaled from terabytes to petabytes without changing their data model—read the full story. https://tsdb.co/xfbkqubo

Four out of five Tiger Hunt tasks you can do at 11pm on a Tuesday. Task 5 is a workshop, and those have dates on them. Plan around it. Get the swag. https://lnkd.in/ey8wetQk

The first Tiger Hunt boxes are going out. Tee, hat, sticker sheet. Earned, not handed out, which is the part we like. If you haven't started: five hands-on tasks in Tiger Cloud. Install the Tiger CLI or MCP, create a hypertable, load data, create a continuous aggregate, join a workshop. Finish all five and one of these ships to you. Not too late. The hunt is still open. https://lnkd.in/gKjRttTj #earnyourstripes #jointhetigerhunt #TigerHunt2026 #Postgres #TimeSeries

Testing database changes is where teams can feel the risk. Schema updates mean holding your breath until they work. We built Tiger CLI and our MCP Server so you can fork production at zero cost, test all your changes in isolation, and push verified updates back with confidence. Test against real data. See what breaks. Fix it. Deploy. Watch how it works in our latest video. https://lnkd.in/gDQJmrsM #PostgreSQL #TimescaleDB #Database #Infrastructure

An empty database teaches you nothing. Task 3 of Tiger Hunt is load data. It looks like the throwaway step. It isn't. Every interesting thing about a time-series database only shows up once there are rows in it. Compression ratios need data. Query plans need data. A continuous aggregate over an empty table is a lesson in nothing. Two options. Use our quickstart and get a populated instance in about five minutes. Or load your own dataset, which is the better choice, because then every later task gets measured against a workload you actually recognize. Then the rest of the hunt has something to bite into. Five tasks, one swag pack: https://lnkd.in/ey8wetQk #Postgres #TimeSeries #DataEngineering #PostgreSQL

The more I talk to data center and bitcoin mining operators, the more they sound like my oil and gas clients Business model is simple: maximize uptime and maximize efficiency. Everything else is a lever on those two metrics. Both need to instantly answer: -Real-Time Equipment Status: Is this pump/GPU running at spec? -Anomaly Detection: Is this temperature or power draw abnormal? -Instant Alerts: This equipment is degrading. -Historical Baselines: What's normal vs today? -Root Cause Analysis: What went wrong and why? -Efficiency Trends: Is output per unit declining? -Predictive Maintenance: What fails next week? From my discussions with data center operators, their most common challenge is the same as our oil and gas clients - getting answers to these questions quickly and accurately. Most split across three systems. Real-time alerting. Historical analytics. Asset metadata. When a GPU fails, ops teams stitch data from all three places. What caused it? Which cooling strategy failed? Are other assets affected? Root cause takes hours because metadata and time-series are separated. The main project we work on for both industries is consolidating that data Our solution is simple keep everything in Postgres Tiger Data extends Postgres to handle millions of time-series events per second while keeping metadata unified in the same database. Allowing you to store operational metadata and time-series in the same Postgres database. -Take advantage of the guaranteed consistency Postgres offers -Have scaled ingest and query performance -Keep metadata and Time-series data tied together at the DB layer When GPU 847 alerts with a temp spike. One query answers: which rack, which cooling circuit, maintenance history, efficiency baselines, how similar GPUs compare, correlation with cooling strategy. "If uptime and efficiency are your only metrics, but metadata lives separate from time-series, how confident are you in either?" Run through those seven questions. How many can you answer with one query? #postgres #TigerData #Crypto #OFS