For Agents
Provision Cloud Composer environments, execute Airflow CLI commands remotely, take snapshots, and manage user workloads. Lets agents drive managed Apache Airflow without operating the underlying cluster.
Install Jentic One Beta
Jentic One is a self-hosted execution layer for AI agents. It lets your agent call the Cloud Composer API, or any other public or private API you need. You set the rules, the agent never sees your credentials, and every call is logged.
Two steps, two machines. Install the instance in a safe environment, then register your agent from wherever it runs.
Step 1: Jentic One Host machine
# On the machine that will host your Jentic One instance:
curl -fsSL https://raw.githubusercontent.com/jentic/jentic-one/main/tools/install.sh | shStep 2: Agent machine
# On the machine where your agent runs (keep this separate from the instance):
curl -fsSL https://raw.githubusercontent.com/jentic/jentic-one/main/tools/install.sh | sh
jentic register # connects your agent to your Jentic One instanceJentic One is in public beta. The setup above keeps your agent separate from the instance, which is what you want before using real credentials: an agent running as the same OS user as Jentic One can read its stored keys directly. Just evaluating? A single local install is fine to start. See the secure deployment guide for the tiers.
What an agent can do with Cloud Composer API.
Create, update, and delete managed Apache Airflow environments
Execute Airflow CLI commands inside an environment and poll their output
Save and load environment snapshots for backup, upgrade, and migration
GET STARTED
Use for: I need to create a new Cloud Composer environment, Run an Airflow CLI command in my Composer environment, Trigger a database failover for a Composer environment, Take a snapshot of a Composer environment for backup
Not supported: Does not author DAG code, run jobs outside Airflow, or replace data processing engines like Dataflow — use for managing Cloud Composer environments and remotely executing Airflow commands only.
Cloud Composer is Google Cloud's managed Apache Airflow service. The API provisions and manages Composer environments, runs Airflow CLI commands remotely, polls long-running command output, triggers database failover, snapshots an environment, and manages user workloads on the underlying GKE cluster. It is the orchestration backbone for ETL, ML pipelines, and cross-system workflows that need scheduled DAG execution without standing up Airflow infrastructure.
Trigger manual database failover for high-availability environments
Manage Kubernetes user workloads on the environment's GKE cluster
Check upgrade compatibility before rolling a new Composer version
Patterns agents use Cloud Composer API for, with concrete tasks.
★ Managed Airflow For ETL Pipelines
Provision a Composer environment to run Apache Airflow DAGs that orchestrate data extraction, transformation, and loading across BigQuery, Cloud Storage, and external systems. The environments.create endpoint stands up the Airflow web server, scheduler, and workers behind the scenes so platform teams skip the operational overhead of self-hosted Airflow.
Create a Composer environment named acme-prod-etl in us-central1 with node count=3 and machine type=n1-standard-2
Remote Airflow CLI Execution
Trigger DAG runs, pause DAGs, list connections, and manage variables by executing the Airflow CLI remotely against a Composer environment. The executeAirflowCommand endpoint accepts the command and arguments, returns an execution ID, and the pollAirflowCommand endpoint streams output, so CI/CD systems can manage Airflow state without SSHing into a worker.
Call executeAirflowCommand with command=dags trigger, args=[my-etl-dag], poll until exit_code is set, and return the captured stdout
Disaster Recovery Snapshots
Take a Composer environment snapshot before a major upgrade and restore from it if the upgrade fails. The saveSnapshot and loadSnapshot endpoints capture DAGs, plugins, and Airflow metadata into Cloud Storage and replay them into the same or a different environment, giving teams a clean rollback path that does not depend on hand-restoring DAG files.
Call saveSnapshot on projects/acme/locations/us-central1/environments/prod, then call loadSnapshot on a freshly-created environment with the returned snapshot URL
Database Failover For HA Composer
Trigger a manual database failover when an outage in the primary zone affects the Composer metadata database. The databaseFailover endpoint switches the active database to the standby in the secondary zone with no DAG redeploys required, which keeps long-running pipelines available during regional impairments.
Call databaseFailover on projects/acme/locations/us-central1/environments/prod and poll the operation until the failover completes
Agent-Orchestrated Pipeline Runs
An AI agent can trigger DAG runs and inspect their state through Composer's API by going through Jentic. The agent searches for 'trigger airflow dag', loads the executeAirflowCommand schema, and invokes the right CLI without learning Airflow's full command surface.
Use Jentic to call executeAirflowCommand on the prod environment with command=dags trigger and args=[customer-segmentation-dag]
21 endpoints — cloud composer is google cloud's managed apache airflow service.
METHOD
PATH
DESCRIPTION
/v1/{+environment}:executeAirflowCommand
Execute an Airflow CLI command in the environment
/v1/{+environment}:pollAirflowCommand
Poll the status and output of a running Airflow command
/v1/{+environment}:saveSnapshot
Snapshot a Composer environment to Cloud Storage
/v1/{+environment}:loadSnapshot
Restore a Composer environment from a snapshot
/v1/{+environment}:databaseFailover
Trigger a manual database failover
/v1/{+environment}:checkUpgrade
Check upgrade compatibility for an environment
/v1/{+environment}:executeAirflowCommand
Execute an Airflow CLI command in the environment
/v1/{+environment}:pollAirflowCommand
Poll the status and output of a running Airflow command
/v1/{+environment}:saveSnapshot
Snapshot a Composer environment to Cloud Storage
/v1/{+environment}:loadSnapshot
Restore a Composer environment from a snapshot
/v1/{+environment}:databaseFailover
Trigger a manual database failover
Three things that make agents converge on Jentic-routed access.
Credential isolation
Google OAuth 2.0 refresh tokens for Cloud Composer are stored encrypted in the Jentic vault. The agent only sees scoped short-lived access tokens at execution time, and credentials for downstream services Composer uses (BigQuery, GCS) stay inside Composer's own service account.
Intent-based discovery
Agents search Jentic with intents like 'trigger airflow dag' or 'create composer environment' and Jentic returns the matching environments or commands operation along with its request schema.
Time to first call
Direct integration with Cloud Composer: 3-7 days for OAuth setup, environment provisioning, and CLI command plumbing. Through Jentic: under 1 hour — search, load schema, execute against an existing environment.
Alternatives and complements available in the Jentic catalogue.
Specific to using Cloud Composer API through Jentic.
What authentication does the Cloud Composer API use?
It uses Google OAuth 2.0 with the https://www.googleapis.com/auth/cloud-platform scope. Through Jentic the OAuth credentials are stored encrypted in the Jentic vault and the agent only ever sees a short-lived access token at execution time.
Can I trigger an Airflow DAG run with the Composer API?
Yes. Call executeAirflowCommand with command=dags trigger and the DAG ID as the argument, then poll pollAirflowCommand for the result. This is the supported way to drive DAG runs without SSHing into a worker pod.
What are the rate limits for the Cloud Composer API?
Environment-level operations like create, update, and delete are heavily rate-limited per project because each one provisions infrastructure under the hood. Airflow command execution and polling allow much higher rates and are usually limited only by Airflow's scheduler throughput.
How do I run an Airflow command through Jentic?
Search Jentic for 'run airflow command', load the executeAirflowCommand schema, and execute the call against /v1/{environment}:executeAirflowCommand with the command and args fields. Jentic returns the schema so the agent does not need to read the Composer discovery doc.
Is the Cloud Composer API free?
Composer is billed by environment-hour and underlying resource consumption (GKE nodes, Cloud SQL, Cloud Storage). The API itself has no per-call charge but every running environment incurs cost, so dev and prod environments should be sized intentionally.
Can I snapshot and migrate a Composer environment?
Yes. Call saveSnapshot to capture DAGs, plugins, and Airflow metadata into Cloud Storage, then call loadSnapshot on the destination environment to restore. This is the canonical pattern for upgrades and cross-region migrations.
/v1/{+environment}:checkUpgrade
Check upgrade compatibility for an environment