IMD SolutionComputing LibraryShop main store

Comparison

Evaluating Ignition historian with TimescaleDB for your plant

Edge computing and IIoT Published 6 min read

Short answer

The Gold Technology Provider announcement is a reason to test TimescaleDB behind Ignition, not proof it replaces your historian. No footprint, compression or performance number is documented for this pairing. Adopt it only after a replay test on your exact Ignition 8.3 build measures growth, ingestion continuity, query response and backup/restore against your current store.

On this page
  1. What the partnership establishes, and what it does not
  2. Four ways to store Ignition history in 8.3
  3. The documented floor, and what it does not size
  4. Sizing five years of storage from a one-week replay
  5. The acceptance test that gates adoption
  6. The decision, case by case
  7. Questions asked

The Tiger Data announcement is a reason to test TimescaleDB behind Ignition, not evidence that it replaces anything. Inductive Automation naming the company behind TimescaleDB a Gold Technology Provider documents a vendor relationship and nothing more: no supported Ignition configuration, no migration path, and no footprint or compression figure for Ignition historian on TimescaleDB appears in the announcement or on the module pages. The decision comes from a replay test on your exact Ignition 8.3 build that measures database growth, ingestion continuity, query response and backup/restore against your current store. Below: what is documented, what is only field-reported, and the numbers that settle it.

What the partnership establishes, and what it does not #

Gold Technology Provider is a tier in a vendor partner program. It says two companies work together commercially; it is not a product certification, a support commitment or a benchmark. Nothing in the announcement or the module documentation states a supported Ignition and TimescaleDB configuration, a documented migration behavior, or a performance guarantee you could take into a design review.

What exists today is thinner but concrete. The vendor's module showcase lists Mustry TimescaleDB Historian, which extends the Ignition 8.3 system.historian interface, and its release discussion describes configuration following normal historian-provider setup, with existing tag groups, scripts and views continuing to work. Field reports add two impressions: a TimescaleDB-backed store takes much less space than writing straight to a SQL database, and an Ignition historian on PostgreSQL has been running well, with TimescaleDB better suited to time-series work but missing some features of premium options such as Aveva PI, at lower cost.

No report carries a tag count, sample rate, retention window, query load, hardware specification or measurement method. They are reasons to measure, not measurements.

Four ways to store Ignition history in 8.3 #

The comparison that matters is between the paths you can actually deploy today:

Storage path How it reaches Ignition Documented Reported in the field Gate before adoption
SQL historian to PostgreSQL Bundled historian module writing to a SQL database PostgreSQL is a documented store for Ignition historical data Long-running deployments described as working well Your measured baseline; every other path is judged against it
Built-in time-series store Bundled with Ignition 8.3 Positioned by the vendor for two to five years of on-premise history Reported short of some enterprise functionality at release, with the rest on the vendor roadmap Verify the enterprise functions in your build, not on the roadmap
TimescaleDB via module Mustry TimescaleDB Historian from the module showcase, extending system.historian The listing covers Ignition 8.3; no footprint or performance numbers Much smaller footprint than direct SQL storage, no ratio or conditions given Full replay test with versions pinned
Premium enterprise historian Separate product integrated with Ignition Complete enterprise feature set under its own licensing The option with nothing missing, at the highest cost Buy only when a required feature fails in all three paths above

The documented floor, and what it does not size #

The historian module page for Ignition 8.3 states a dual-core processor, 4 GB RAM and 10 GB of free disk space, with Windows Server 2016 and 2019 among the supported operating systems, and adds that requirements vary by usage. That floor has one use: it tells you whether a host can install and run the module at all. It does not size a store, and no source documents a per-tag storage rate for any path in the table. Host specification comes from the replay test.

Placement is the other decision at this level: the database can share the gateway host or get its own box. When it moves to its own box in the cabinet, a fanless mini PC carries the test image and, on a pass, the production store, so test and production run on one image and hardware stops being a variable. The same measure-then-size method applies to other self-hosted stacks; sizing a self-hosted Node-RED InfluxDB Grafana historian walks the InfluxDB version of it.

Sizing five years of storage from a one-week replay #

Inputs, with units and sources:

  • Retention horizon: five years on-premise, the upper figure the vendor names for built-in history; two years is the lower.
  • Install floor: 10 GB of free disk, from the historian module page for Ignition 8.3.
  • Measured growth: from your own seven-day replay of the intended tag volume; the 20 GB per week used here is an illustrative stand-in for that measurement, because no documented rate exists.

Rule in words: projected store size equals weekly growth times the weeks in the horizon, plus the install floor; the volume you provision must meet or exceed that sum at the horizon you choose.

Arithmetic, written out:

  • Five years is 5 × 52 = 260 weeks.
  • 20 GB/week × 260 weeks = 5,200 GB, which is 5.2 TB; adding the 10 GB floor gives 5,210 GB to provision.
  • Two years is 2 × 52 = 104 weeks: 20 GB/week × 104 weeks = 2,080 GB, about 2.1 TB, plus the same floor.

Decision: if the volume the historian host can carry is below the projection plus floor, the configuration fails on capacity at that horizon. Drop to the two-year figure, lower the weekly rate by sampling fewer tags less often, or change the host. With the projection met and a second replay week confirming the rate, disk is settled before any other judgment is made.

The acceptance test that gates adoption #

  1. Pin the versions: the exact Ignition 8.3 build, the module build from the showcase listing, and the database release, all recorded in the test log. Every result below is valid only for that set.
  2. Replay the intended tag volume and update pattern for seven days on production-like hardware, on a host separate from the production gateway so a test fault cannot take running history with it. The week yields the growth rate used for sizing.
  3. Check ingestion continuity: compare samples expected against samples stored across the window. Any unexplained gap fails the run.
  4. Measure query response with the views operators actually open. The pass mark is that every trend loads inside its own refresh interval on the test host.
  5. Exercise backup, restore, restart and failure recovery, and confirm the restored history is complete and queryable from Ignition before passing.
  6. Test each enterprise function you assume, such as redundancy, administration and reporting interfaces, directly, because only a function that works in your build counts, whatever the roadmap says.
  7. Run the identical workload against the incumbent store and compare the two logs from the same hardware.

One measurement deserves a second look before it condemns the store: on TimescaleDB with compression enabled, the weekly rate already includes compression's effect, and an unexpectedly high rate can mean compression never engaged. The failures covered in TimescaleDB hypertable compression not working appear as fast growth with no error on the Ignition side.

The decision, case by case #

  • PostgreSQL historian running today with nothing missing: keep it. Move only if the replay shows materially lower growth with equal continuity, query and restore results on your build.
  • Two to five years of on-premise retention without premium licensing: test the built-in time-series store's enterprise gaps first. It is the bundled, documented path for exactly that horizon.
  • TimescaleDB module: adopt only on a passed test with versions pinned in the log. The Gold Technology Provider announcement is not a support contract for it.
  • A required feature that fails in all three paths: the premium historian is the answer, and its higher cost is the trade the field reports already describe.

If the open question is which historian product to run at all, rather than which storage engine sits under Ignition, start from plant historian selection by scale and support needs.

Questions asked

Does Gold Technology Provider status mean Inductive Automation supports TimescaleDB?

No. It is a partner-program tier describing a vendor relationship, not a product certification or a support commitment for any configuration. Support for the TimescaleDB path comes from the module that provides it, so before a production move you confirm that module vendor's version matrix covers your exact Ignition 8.3 build and that its support channel meets your response-time needs.

How much smaller is the TimescaleDB footprint than PostgreSQL for Ignition?

One field report says much smaller, with no tag count, sample rate, retention window, hardware or measurement method behind it, so no percentage transfers to your plant. Measure it: run the same replay against both stores in the acceptance test and divide the two growth rates. That ratio, measured on your build, is the only footprint number you should size with.

Is the built-in 8.3 time-series historian ready to replace an enterprise historian?

A field report says the QuestDB-based option in 8.3 was not enterprise-ready at release because required functionality was not all implemented, with the remainder reported to be on the vendor roadmap, and a roadmap statement is not a delivered function. Gate it on the same acceptance test, with extra weight on the enterprise items: backup, restore, redundancy and administration in your exact build.

What host does the Ignition historian module need?

The documented floor for Ignition 8.3 is a dual-core processor, 4 GB RAM and 10 GB of free disk, with Windows Server 2016 and 2019 among the supported operating systems, and the module page itself says requirements vary by usage. The floor proves an install can run; it does not size your store. Your measured CPU and RAM use during the replay, plus the disk projection, are the real host specification.

Do existing tags, screens and scripts keep working with a TimescaleDB module?

The module's release discussion reports that setup follows normal historian-provider configuration and that existing tag groups, scripts and views keep working. What no source documents is migration of already-stored history into the new store. Treat that as untested until you restore a backup into it and query the old date range from Ignition with stored counts matching the source.

Keep reading

  • How-to

    Backfilling PI from an existing plant historian

    Split the historian-to-PI backfill into four independent stages — extract, intermediate store, external compression, destination load — and qualify throughput with a measured pilot. Rows divided by elapsed seconds, not tag count, sets the schedule: 12,000 tags from 2020 at 500 rows/s is about 203 days. Check the native Historian 2023 replication route before writing pipeline code.

    Edge computing and IIoT

  • How-to

    Backfilling trend history from a plant historian

    Yes, you can backfill trend history from a plant historian into a SCADA server. Extract the stored records through the historian's native export or ODBC interface, map each source point to its destination object with original timestamps and quality, insert the batches through the destination's documented history methods, and prove the import by reading the interval back from the destination. A saved CSV is a bridge, not proof of import.

    Edge computing and IIoT

  • Selection guide

    Plant historian selection by scale and support needs

    Pick the tier first: a control-room or automation-vendor historian for one site's process trends, an enterprise platform only when several sites share one operational history, and a self-built time-series stack only when your team owns connectors, dashboards, backups and upgrades. Size by stored events per second — the sum of each tag group's tags times its stored events per second — times retention, never by tag count alone.

    Edge computing and IIoT