New module: TimescaleDB historian for Ignition 8.3

Hi all,

We've released a TimescaleDB-backed historian module for Ignition 8.3 that I wanted to share with the community, in case anyone is running into scaling limits with the native historian.

The short version: it extends system.historian with TimescaleDB, so you get fast, scalable historian workloads without giving up the native Ignition tag and query experience. It's configured like any other historian provider, and your existing tag groups, scripts and views keep working.

Why it's better then the old SQL Historian:

  • Database-side aggregation. Default aggregation is pushed down into TimescaleDB, so queries stay fast even at millions of samples per tag.
  • Tag renames without SQL. Rename history tags from inside Ignition, no touching the underlying tables.
  • Annotations and metadata. Attach rich context to historian data so it travels with the values.
  • Built for scale. Aimed at sites where the default historian starts to struggle: millions of tags, long retention, mixed analytical workloads.

On performance, we ran the same test across setups: a 1000-point aggregation of a single tag on a ~158M-row dataset (queryAggregatedPoints, Average mode, median of 5 runs). Our module stayed flat in the tens of milliseconds from a week up to a full year of data, while the SQL historians climbed past 35 seconds on the year query and the Core Historian ran out of memory. Rough numbers:

Window Our module SQL (TimescaleDB) SQL (Postgres) Core Historian
Week 25 ms 489 ms 493 ms 7.7 s
Month 38 ms 2.7 s 3.1 s 12.3 s
Year 56 ms 37.3 s 35.8 s Out of memory

It was built by our team out of real historian projects and is supported by people running it in production. Pricing is €3,000 per module, with optional updates through a care module. System integrators get a discount, and there's a bulk discount for larger single orders.

Details, benchmarks and download here: TimescaleDB Module for Ignition 8.3 | Mustry Solutions

Happy to answer any questions about setup, benchmarks, or how it behaves alongside the native historian.

Do you know this compares to Flow software timebase historian?

Would be intesting to see.

I don't know much about Timebase tbh, but I would think that Timebase is not Open Source, and does not have a large community that works on it, nor do a lot of people (or AI) know very well how to do advanced analytics on it.

TimescaleDB is a PostgreSQL extension, so you get full SQL, PostgreSQL's ecosystem, joins, and relational features on top of time-series capabilities. That's something I like about it at least.