Astro Features
Everything you need to build, run, and observe production data and AI pipelines, organized by the capabilities that matter most to your workflow.
Build
Ship production-ready pipelines faster.
Write Dags with an AI that actually understands Airflow. Test against production-parity environments in your browser. Deploy with a single command click.
new
Astro IDEAstro IDE is a browser-based development environment built specifically for Airflow. Unlike general-purpose AI coding assistants, Astro IDE understands your data stack—your Airflow version, operators, and Dag patterns.
The built-in AI generates production-ready code using modern syntax and modular structures. Validate Dags against production-parity environments without installing Docker locally—ephemeral test deployments spin up on demand in seconds with the same runtime and connections your production Dags use and automatically shut down when idle.
Development that used to take an hour now takes 15 minutes. Go from natural language prompt to running Dag without leaving your browser.
new
Astronomer's AI Agent ToolingOpen-source AI agent tooling maintained by Astronomer that brings Airflow expertise to local development environments. Install it in Claude Code, Cursor, VS Code, or 25+ other AI coding tools to get specialized skills for Dag authoring, debugging, Airflow 2→3 migration, lineage tracing, and warehouse operations.
Unlike generic AI assistants that hallucinate outdated patterns, the agent tooling understands Airflow versions, operator patterns, and best practices. It accesses real execution data for debugging (task logs, run history, and failure patterns) instead of guessing at solutions. Context-aware recommendations adapt to your specific Airflow version and installed providers.
Development stays in your preferred editor. Write, test, and debug production-ready Dags without switching tools or leaving your local workflow.
Astro CLI
The Astro CLI provides a complete local development experience for teams that prefer working in their own editors. Run astro dev start to spin up a local Airflow environment with hot-reloading—the Airflow UI automatically opens when the webserver is ready, and your Dag changes reflect immediately without manual restarts.
Deploy with astro deploy, which builds your project image, authenticates to Astro, and pushes to your deployment in a single command. The CLI also supports astro dev parse for quick Dag validation and astro dev pytest for running your test suite before deploying.
CI/CD & Preview Environments
Preview Deployments let you test Dag changes in isolated environments before merging to production. When you create a feature branch, Astro can automatically create a corresponding preview deployment that mirrors your production configuration—same Airflow version, same connections, same environment variables.
Astro provides pre-built CI/CD templates for GitHub Actions, GitLab CI, Azure DevOps, AWS CodeBuild, Bitbucket, CircleCI, and Jenkins. The GitHub deploy-action automatically chooses deploy types based on which files changed—Dag-only deploys when you update files in the dags directory, full image deploys when you modify other project files.
Run your test suite before deploying and block merges if tests fail. For teams using dbt, deploy templates bundle your dbt project files and push them to Astro where they're accessible to Dags using Cosmos.
Connection & Environment Management
Centrally manage Airflow connections and environment variables across all deployments. Create connections once and link them to multiple deployments—no need to recreate credentials in each environment.
Store sensitive values securely and control access with workspace-level permissions. Update connection details in one place and have changes propagate to all linked deployments automatically.
Astro Terraform Provider
Manage Astro infrastructure as code with the official Terraform provider. Define deployments, workspaces, API tokens, and environment variables in version-controlled Terraform configurations.
Automate provisioning and changes across environments while maintaining consistency and auditability. Integrate Astro into your existing Infrastructure as Code workflows.
dbt Integration with Cosmos
Cosmos is an open-source framework maintained by Astronomer that turns dbt projects into Airflow Dags with model-level visibility. Instead of running your entire dbt project as a single task, Cosmos creates individual Airflow tasks for each dbt model, test, seed, and snapshot.
This gives you granular control over retries, dependencies, and resource allocation at the model level. When a model fails, you can rerun just that model and its downstream dependencies—not your entire dbt project. Lineage flows through from dbt models to Airflow tasks to downstream consumers, giving you end-to-end visibility in Astro.
Astro supports deploying dbt code directly alongside your Dags, simplifying collaboration between dbt developers and Airflow engineers working on the same pipelines.
Run
Scale to any workload without managing infrastructure.
The Astro engine delivers 2x the performance of other Airflow solutions. Auto-scaling, high availability, and multi-cloud deployment, all made easy for you.
Astro Executor
The Astro Executor is a new execution model exclusive to Astro that eliminates the reliability issues of Celery and the cold-start issues of Kubernetes executors. Instead of queueing tasks through external message brokers like Redis or RabbitMQ, lightweight agents communicate directly with the Airflow API server.
This architecture provides predictable performance—tasks start executing immediately when agents are available, with no waiting for broker polling or pod startup delays. It also delivers superior reliability during infrastructure disruptions: while Celery workers can lose connection and orphan queued tasks during scaling events or broker restarts, the Astro Executor maintains persistent connections that prevent single points of failure.
In extended testing, the Astro Executor maintained zero task failures while Celery experienced occasional task and Dag failures under identical conditions. The direct agent communication model also enables coordinated scale-down that avoids terminating actively working agents during autoscaling events.
Elastic Auto-Scaling
Workers scale automatically based on task queue depth. When the number of queued and running tasks exceeds your configured concurrency, Astro adds workers to handle the load. When demand drops, workers scale back down, including scaling to zero for worker queues that don't need always-on capacity.
You configure min/max worker counts and concurrency settings per worker queue, giving you control over scaling behavior while Astro handles the infrastructure. This means you pay for compute only when tasks are actually running, not for idle workers waiting for work.
Task-Optimized Worker Queues
Configurable worker queues let you create optimized execution environments for different types of tasks within the same deployment. Separate resource-intensive ML tasks from lightweight SQL queries. Isolate long-running jobs from short tasks. Configure some queues to scale to zero while keeping others always available.
Each queue can have different worker types (CPU/memory configurations), concurrency settings, and scaling behavior. Assign tasks to queues in your Dag code with a simple queue='queue-name' argument, and Astro routes the task to the appropriate workers.
Astro offers the largest worker compute options in the managed Airflow market—up to 8x larger than alternatives—for tasks that need significant resources.
Event-driven Scheduling
Airflow 3 introduces event-driven scheduling that lets pipelines react to data changes or external triggers in near real-time, rather than relying solely on cron-based schedules. Integrate with message systems like Amazon SQS to trigger Dags the moment data arrives.
This capability is essential for use cases where latency matters—processing streaming data, responding to user events, or triggering inference pipelines when new inputs are available. Combined with the removal of the uniqueness constraint on logical dates, you can now run multiple instances of the same Dag simultaneously for parallel inference workloads.
Remote Execution
Remote Execution Agents let you run tasks in your own infrastructure—on-premises data centers, private clouds, GPUs, or environments with strict network controls—while Astro handles orchestration. Agents establish only outbound, encrypted connections to Astro's Orchestration Plane, eliminating the need to open inbound firewall ports.
Sensitive data never leaves your infrastructure. Code stays in your secure repositories. Secrets remain in your environment. You get the benefits of managed orchestration—scheduling, UI, observability, autoscaling—without compromising on security or compliance requirements like HIPAA, GDPR, or SOC 2.
You can run agents in multiple regions or cloud environments simultaneously. If one environment goes down, another can automatically pick up the work without manual failover.
High Availability
Astro runs Airflow components across multiple availability zones to prevent single points of failure. If a node or availability zone has issues, Dags continue running without interruption. Schedulers, webservers, and triggerers are all configured for redundancy.
Astro provides a 99.5% uptime SLA for production workloads, backed by 24/7 support with 1-hour response times on enterprise plans. Deployment rollbacks let you quickly revert to a stable state during upgrades or when accidental configuration changes cause issues.
Multi-cloud Deployment
Deploy Astro on AWS, Azure, or Google Cloud in 55+ regions worldwide. Choose the region closest to your data sources to minimize latency and meet data residency requirements. Astro is available through AWS Marketplace, Azure Marketplace, and Google Cloud Marketplace, letting you use committed cloud spend toward your subscription.
For teams with strict network requirements, Astro Private Cloud provides dedicated infrastructure within your own cloud environment while Astronomer manages the control plane.
Inference Execution
Airflow 3 removes the single "execution date" constraint that previously prevented running multiple instances of a Dag simultaneously. This is purpose-built for GenAI pipelines, large-scale predictions, and on-demand model serving where you need to process multiple inference requests concurrently.
Trigger Dags via API calls, launch parallel runs for experimentation, and process inference workloads dynamically without being constrained by predefined schedules. Combined with event-driven scheduling, you can build near real-time inference pipelines that respond to incoming requests as they arrive.
Scheduler-managed Backfills
Backfills—reprocessing historical or newly available data—were previously triggered from command-line processes that could terminate if the session was lost, leaving long reruns vulnerable to interruption. In Airflow 3, backfills are first-class citizens managed by the scheduler itself.
Trigger backfills asynchronously via the UI or API, monitor progress in real-time, and pause or cancel jobs mid-run. Large-scale historical recalculations that take hours or days to complete now run consistently and reliably, which is critical for ML retraining and data integrity checks.
Observe
Detect issues before they impact downstream systems.
Real-time lineage, SLA monitoring, and automated RCA, all built into the orchestration layer, not bolted on after the fact.
Real-Time Pipeline Lineage
Astro automatically extracts lineage metadata from Airflow tasks as pipelines execute using OpenLineage. View upstream and downstream dependencies in an interactive graph that updates in real-time.
Connect your Snowflake or Databricks warehouse to extend lineage visibility to table-level dependencies—including tables not directly touched by Airflow tasks. This gives you complete end-to-end visibility from source systems through transformations to final warehouse tables.
Quickly assess blast radius when issues occur—trace failures to upstream root causes or identify all downstream assets affected by a change. Lineage spans Dags, deployments, and warehouse tables, providing complete visibility across your data platform.
Data Product SLA Tracking
Define service level agreements for your most important data products and track whether they meet business requirements. Set timeliness SLAs for delivery by specific times on selected days (e.g., every day by 9am EST), freshness SLAs for required update intervals (e.g., never more than 2 hours stale), or custom SLAs with full control using cron expressions.
Astro evaluates SLAs by checking successful runs of final assets in your data product against your defined parameters. The Data Products dashboard provides a unified view for business stakeholders to see which products are late or stale, who owns them, and historical SLA compliance at a glance.
Predictive Alerting
Go beyond reactive alerting with proactive alerts that notify you before failures disrupt delivery. Proactive SLA Alerts monitor upstream dependencies of your data products and notify you when delays in upstream tasks risk causing SLA misses—not just after they're already missed.
Proactive Failure Monitors alert you when any upstream or final Dag in your data product fails. Alerts include direct links to lineage views showing the failing Dag and task, plus identification of all downstream assets impacted by the failure—surfacing both root cause and blast radius in a single view.
Integrate alerts with Slack, PagerDuty, email, and webhooks for routing to your existing incident response workflows.
AI-Assisted RCA
Accelerate root cause analysis with AI assistance that helps you diagnose and resolve failures faster. When tasks fail, Astro generates AI-powered summaries that highlight what broke, why, and how to fix it—without requiring you to sift through Airflow logs.
Combined with real-time lineage visualization, you can quickly understand the blast radius of failures and trace issues back to their source. This dramatically reduces troubleshooting time, especially for teams managing many Dags or onboarding new engineers who aren't yet familiar with the codebase.
Data Quality Monitoring
Monitor data quality in your warehouse with built-in checks or custom SQL monitors. Track column null percentages, row volume changes, and schema drift with pre-built monitors—or define custom SQL-based checks tailored to your specific requirements.
Run monitors on scheduled intervals or trigger them based on pipeline events for real-time validation as data flows through your pipelines. When a quality issue is detected, trace it back to the upstream Airflow task that caused it—monitors are linked directly to pipeline execution with table-level visibility that extends beyond what Airflow orchestrates for faster resolution with full context.
Asset Catalog
The Asset Catalog automatically captures all pipeline assets, datasets, and tables for complete visibility across your data platform. Browse assets by namespace, filter by source system, and see ownership information for each asset.
This provides a single source of truth for what exists in your data ecosystem, who owns it, and when it was last updated—making it easier to track down the right person when issues arise.
Health Dashboard
Monitor data product health across all deployments from a unified dashboard. See which products are late or stale, who owns them, and when they were last updated—all at a glance.
Track operational metrics like Dag runtimes, task failures, and SLA compliance across your entire Airflow ecosystem. For platform teams managing multiple deployments, this centralized view provides visibility that open source Airflow alone cannot deliver.
Enterprise
Security and governance without compromise.
Astro is built for organizations with the most stringent data security requirements—comprehensive compliance certifications, fine-grained access control, and enterprise-grade encryption.
Compliance Certifications
Astro is compliant with SOC 2 Type II controls for security, availability, and confidentiality. Astronomer is GDPR-compliant as an organization and offers Data Processing Agreements (DPAs) that satisfy GDPR requirements for data processors.
For organizations handling protected health information, Astronomer is HIPAA-compliant and Astro is HIPAA-ready—available through a Business Associate Agreement (BAA). For payment card data, Astro is certified PCI DSS compliant.
Visit the Astronomer Trust Center to request SOC 2 Type II reports, penetration test reports, and other compliance documentation.
Single Sign-On & SCIM
Integrate Astro with your identity provider for secure single sign-on (SSO) authentication. Astro supports Okta, Microsoft Entra ID (Azure AD), and other SAML 2.0 and OIDC-compatible providers. Multi-factor authentication (MFA) adds an additional verification layer beyond passwords.
SCIM (System for Cross-domain Identity Management) integration automates user provisioning and deprovisioning. When employees join or leave your organization, their Astro access is updated automatically based on your identity provider configuration.
Role-Based Access Control
Astro provides fine-grained role-based access control (RBAC) at the Organization, Workspace, and Deployment levels. Assign precise permissions that restrict access to specific resources, aligning with the principle of least privilege.
Custom Deployment Roles let you create roles tailored to your organization's needs. For example, create a read-only role for stakeholders who need visibility without edit access, or a restricted operator role that can trigger Dags but not modify configurations.
Deployment API tokens can be scoped to specific deployments and assigned limited permissions, making them safer for CI/CD pipelines than broad-access tokens.
Encryption
All data is encrypted in transit using TLS 1.3 with strong ciphers (TLS 1.2 available upon request). Data at rest is encrypted using keys stored in AWS KMS, Azure Key Vault, or Google Cloud KMS depending on your deployment region.
For additional control, you can use customer-managed encryption keys. Connections and environment variables marked as secrets are encrypted with an additional layer of protection.
Network Isolation
Astro supports VPC peering and AWS PrivateLink for private network connectivity from your applications to Astro without traversing the public internet. Configure private endpoints to access data sources in your VPC without exposing them externally.
IP allowlists restrict access to the Astro UI and API to known IP ranges. For teams with strict network security policies, Remote Execution Agents establish only outbound connections, eliminating the need to open inbound firewall rules.
Secret Management
Integrate Astro with your existing secrets backend—HashiCorp Vault, AWS Secrets Manager, Azure Key Vault, or Google Secret Manager—to retrieve connections and variables at runtime. Secrets are never stored in Airflow's metadata database or exposed in logs.
Environment variables can be marked as secret in the Astro UI, encrypting them at rest and masking them in the interface. Customer Managed Workload Identity lets you use existing cloud IAM roles across multiple deployments without duplicating credentials.
Audit Logging
Astro maintains comprehensive audit logs of user actions, API calls, and system events. Track who accessed what resources, when deployments were modified, and which users triggered Dag runs.
Audit logs support compliance requirements for regulated industries and provide the documentation needed for security reviews and incident investigations.
Deployment Rollbacks
Quickly revert a deployment to a previous stable state when issues arise. Rollbacks restore your Dag code, environment variables, and configurations to a known-good snapshot, minimizing downtime during incidents or failed upgrades.
This provides an additional safety net for production deployments, letting teams deploy with confidence knowing they can recover quickly if something goes wrong.
Get started free.
OR
By proceeding you agree to our Privacy Policy, our Website Terms and to receive emails from Astronomer.