Cookies and Privacy

We use technology on our website to collect information that helps us enhance your experience and understand what information is most useful to visitors.
By clicking "I ACCEPT," you agree to the terms of our privacy policy.

Cirata applies the highest standards to its use of data and its compliance with data-protection regulations across our marketing and website. Our Data Protection Officer can be contacted at DPO@cirata.com. You can change your cookie settings at any time via the link in our footer.

Cookie Setting

Orchestration extension

« Back to full extensions directory

Orchestration extension

A vendor-neutral control plane for workflow orchestration on Cirata Symphony. Orchestration connects to the workflow engines you already run, behind one unified Orchestrator interface.

Built by Cirata | Tech Preview
get access schedule 1:1 demo watch technical demos

Built for data orchestration

Orchestration is the layer that federates the engines you already run and reacts to data-platform events across all of them.


Federates:
Apache Airflow (2.x and 3.x, auto-detected) · Dagster · Prefect, behind one Orchestrator contract.


Triggers on:
Messaging subjects · JetStream KV key changes · any sibling Cirata Symphony extension on a poll, Canon, Ice Flow, Pulse, and more.

Built for data orchestration

What it does

  • Federate Apache Airflow, Dagster, and Prefect behind one unified interface, discover and launch workflows across all three from a single surface.
  • Launch a workflow on any connected engine from an event on any other system, the cross-engine trigger is the product, not a connector.
  • Define triggers declaratively with JSONPath rules, no Lambda, no cron, no glue code to own and forget.
  • Fire exactly once on the edge, a condition that stays true launches the workflow once, not on every poll.
  • Keep your orchestrators where they are, Orchestration connects over each engine's existing API, so no re-platforming and no DAG rewrite.
What it does screenshot 1
What it does screenshot 2

Use cases

Cross-engine event reaction

Wire an Iceberg compaction event to an Airflow DAG, or a Kafka offset stall to a Prefect flow, declaratively, in the UI. The single most common piece of unowned glue in a data platform becomes a first-class object.

One pane across every engine

See and launch workflows across Airflow, Dagster, and Prefect from a single interface, with each workflow's inputs rendered as a form regardless of which engine it lives on. No more switching between three vendor UIs.

Retire the bespoke glue

Replace the Lambda watching an S3 bucket, the cron job polling a GraphQL endpoint, and the Slack bot kicking a DAG, all the unmonitored scripts owned by whoever wrote them last, with declarative triggers anyone on the team can see.

Federate without re-platforming

Connect over each engine's existing REST or GraphQL API. Your DAGs, flows, and jobs stay exactly where they are, Orchestration sits above them, not in their place.

How it works

Connect

Register your Airflow, Dagster, and Prefect engines. Orchestration auto-detects versions and surfaces every workflow behind one interface.

Trigger

Define a trigger that watches a messaging subject, a KV key, or a sibling extension, and evaluate incoming events against JSONPath rules.

Launch

When the condition first becomes true, Orchestration extracts parameters from the event and launches the named workflow on the connected engine, exactly once.

Technical details

Federation One Orchestrator interface, list, detail, schema, trigger, over Apache Airflow (REST, 2.x/3.x auto-detected), Dagster (GraphQL), and Prefect (REST deployments).
Trigger sources Messaging subjects, JetStream KV key watches, and sibling-extension polls, any data-platform event on the Symphony bus.
Rules Declarative JSONPath evaluation (equals, contains, starts_with, exists) with all-must-match semantics; parameters extracted from the event or set as static literals.
Trigger semantics Edge-triggered state machine, fires actions once on the WAITING → READY transition; explicit reset to re-arm.
Sources S3-backed workflow definition files, browsable and editable in place; bundled DAG templates to seed new authoring.
Operator surface React UI, 19 messaging endpoints, and MCP / OpenAPI discovery via attached schemas; configuration and trigger state in JetStream KV, no external database.

Other Cirata Symphony extensions

Cirata Symphony Pulse extension
Control, understand, coordinate, and automate your data estate using simple prompts.
view details »
Data Migrator extension
Fully automated, zero-downtime or risk, continuous petabyte scale data migration between data environments.
view details »
Observability extension
See your entire Cirata Symphony estate as one OpenTelemetry stream, every extension, every signal, into any backend you already run.
view details »
Intelligence extension
Connect your data estate to any AI model, without writing a line of integration code.
view details »
Ice Flow extension
Manage data in Iceberg-native, open standard formats, between any pair of catalogs, across vendors, clouds, and on-premises, without compromise or lock-in.
view details »
Canon extension
Replicate Kafka across any vendor mix, with offsets every instance agrees on, schemas that travel with records, and failover you can actually rehearse.
view details »

Want to see the Cirata Symphony Orchestration extension in action?

Let us show you how the Cirata Symphony Orchestration extension is a revolutionary new data extension to fit your organizational needs.
Schedule a direct 1:1 demo with our CTO, Paul Scott-Murphy.
add extension