Tiger Cloud Updates: Bigger Gains, Broader Reach, Clearer Insight

We’ve spent the last few months digging into the three things that matter most to anyone running time‑series workloads on Tiger Cloud – raw performance, how far the platform can reach, and how much insight you get into what’s really happening inside the database.

First up, performance. In the past, squeezing data into a compressed format felt like a trade‑off: you saved space but paid a penalty when you tried to read or write. That mindset is finally fading. With TimescaleDB v2.26, aggregate queries such as COUNT, MIN, MAX or FIRST/LAST now peek straight at the compressed metadata. In real‑world tests that translates into up to 70× faster reads. Even something as common as time_bucket() grouping is about 3.5× quicker.

But it isn’t just reads that get a boost. The newer v2.27 release lets UPDATE, DELETE and UPSERT operations skip the decompression step when the target rows can’t possibly match. The net effect? Selective writes can be up to 160× faster – a genuinely noticeable difference when you’re streaming millions of points per second.

Continuous aggregates, the workhorse for many dashboards, also got a stability makeover. We squashed a memory leak, tightened edge‑case query correctness and removed a deadlock that could appear during concurrent refreshes. The result is a smoother, more reliable experience, especially when you push those aggregates to their limits.

Search lovers, listen up. The pg_textsearch extension (v1.0.0) now brings production‑ready BM25 ranking right into PostgreSQL. No more hunting down a separate Elasticsearch or OpenSearch cluster. In our benchmark of 138 million documents, pg_textsearch was up to 6.5× faster on typical multi‑word queries and handled 8.7× more concurrent traffic than ParadeDB. It even supports the new <@> syntax and a bm25_force_merge() helper for segment consolidation.

On the storage side, Tiger Cloud’s on‑demand scaling has been broadened. You can now pick between 16 k and 40 k IOPS (up to 1.5 GB/s) on regular plans, and up to 80 k IOPS with 2 GB/s and a whopping 64 TB of capacity on enterprise tiers. Adjustments happen without downtime and you only pay for the IOPS you actually consume – no more guessing at a peak that might never arrive.

We also tried to win back the time you’d otherwise spend wrestling with configuration. The console now lets you spin up a hypertable with a few clicks, columnstore‑enabled, and automatically tunes chunk intervals for you. No more hand‑crafting CREATE TABLE statements just to get the right partitioning.

Moving data into Tiger Cloud is smoother, too. The PostgreSQL Source Connector has graduated to production‑ready status. It can replicate an existing PostgreSQL database straight into the cloud, supports configurable worker counts, SSH tunneling and bulk schema mapping. In short – you get a reliable migration pipeline without writing custom code.

Reliability and visibility aren’t just buzzwords; they’re baked in now. We added two Azure regions – Germany West Central (Frankfurt) and Southeast Asia (Singapore) – so teams with GDPR or local‑latency requirements can keep their data where it belongs. Private endpoint support means your traffic can stay off the public internet, and the platform handles disaster recovery, network isolation and telemetry for you out of the box.

All of these pieces together mean you can stay on PostgreSQL as you grow, spend less time fiddling with knobs, and finally get the kind of observability that time‑series workloads deserve. That’s the essence of what’s new in Tiger Cloud.