TigerData review

TigerData is the company behind TimescaleDB, a PostgreSQL database for storing and querying live and historical sensor and machine data. The product is aimed at time-series workloads, but its materials also highlight transactional applications, analytics, events and vector search. It is a managed database service, with compute and storage metered hourly and billed monthly in arrears, according to the pricing page.

The core idea is to keep familiar SQL and PostgreSQL while adding time-series-oriented storage and query features. TigerData says it automatically partitions data by time and key, and uses row storage for writes and columnar storage for analytics, with compression. Its listed capabilities also include tiered storage, more than 200 time-series SQL functions, Kafka and S3 ingestion, Iceberg replication, and hybrid keyword-and-vector search with filters and ranking. Those features make it more than a basic hosted Postgres instance, but the right fit depends on workload and metered cost.

TigerData’s clearest strength is combining PostgreSQL with features designed for time-series data. Automatic partitioning by time and key targets the read and write patterns common in machine data. Row-columnar storage is intended to support both writes and analytics, while compression and tiered storage give teams options for older, colder data. The product materials list more than 200 time-series SQL functions, plus Kafka and S3 ingestion and replication to Iceberg.

The feature set also reaches beyond classic time-series use. TigerData lists hybrid search that combines keyword and vector retrieval with filters and ranking, alongside production security features on its homepage. That breadth may suit teams that want to keep several data patterns in a PostgreSQL-based system rather than assemble separate tools; the materials alone do not establish that every workload should be consolidated this way.

The listed Performance plan starts at $30 per month and is positioned for new applications and internal tools. It allows up to four database services, each with up to 8 CPU, 32 GB memory and 16 TB disk, plus one HA replica and three-day point-in-time recovery. A 30-day trial is listed, though there is no free plan.

The headline plan price is not the whole bill. Compute and storage are metered hourly, billed monthly in arrears, and annual billing is not shown. Performance storage is listed at $0.177/GB-month; Scale starts at $36 per month and lists storage at $0.212/GB-month. Scale allows up to 32 CPU, 128 GB memory and 64 TB disk per service, with up to two HA replicas and 14-day point-in-time recovery. Tiered storage is $0.021/GB-month, but is listed for Scale and Enterprise, not Performance.

Enterprise pricing is not published. Nor are prices for possible IO Boost, production support or VPC peering charges. HA and read replicas are billed at the primary service’s compute and storage rates, so adding availability can add material cost. Enterprise lists up to 64 CPU, 256 GB memory and 64 TB disk per service, but those caps do not come with a public price.

Cost concerns show up in user discussions, but they are not comparable quotes: one Reddit poster reported a $400 monthly bill, while another asked about a $600–$800 setup. Check your own expected compute, storage and replica use before switching.

The listed starting prices are $30 per month for Performance and $36 for Scale. That’s a $6, or 20%, jump to Scale, which raises the per-service ceilings from 8 to 32 CPUs, 32 to 128 GB of memory, and 16 to 64 TB of disk. Enterprise pricing is not published.

Treat those prices as a starting point, not a fixed monthly total: compute and storage are metered hourly and billed monthly in arrears. Storage is listed at $0.177 per GB-month on Performance and $0.212 on Scale. Tiered storage costs $0.021 per GB-month and is listed for Scale and Enterprise. HA and read replicas are billed at the primary service’s compute and storage rates; IO Boost, production support, and VPC peering may add charges, with prices not published.

There’s no free plan, but the pricing details list a 30-day trial. Annual billing is not shown. Model your expected compute, storage, and replicas before comparing the total with a fixed-price alternative.

TigerData is built around Postgres and TimescaleDB, but the published materials don’t specify migration tooling, supported source databases, compatibility limits, or an export procedure. That leaves migration effort and exit steps unconfirmed. Before switching, ask TigerData for a migration path and a tested way to export your data, including any time-series or search features your application depends on.

The product highlights time-based partitioning, tiered storage, SQL analytics functions, and hybrid search. Those features may shape your application’s dependencies; confirm which data and behavior remain portable if you leave. The published materials don’t describe those dependencies or document a rollback process.

We also don’t have Reddit posts or comments in the published material, so we can’t credibly report why users say they leave. One cost question is clear from the pricing details: compute and storage are metered, and replica charges follow the primary service’s rates. Forecast those charges against your workload before committing.