Etlworks offers several deployment models to fit different infrastructure, compliance, and operational requirements. This page walks through the options, the trade-offs, and the questions worth thinking through before deciding.
Deployment Models
Fully Managed SaaS (Cloud)
Etlworks hosts and manages the instance for you. Two flavors:
- Shared Instance — Etlworks-hosted, with resources shared across tenants. Lowest setup overhead.
- Dedicated Instance — Etlworks-hosted, but isolated to your organization. Better for higher throughput, performance isolation, or stricter tenancy requirements.
Best fit when you want to minimize infrastructure management and don't have specific data residency or network isolation requirements.
Customer-Managed Deployment
You host Etlworks in your own environment.
- On-Premise — Etlworks runs inside your data center or private network.
- Customer-Hosted Cloud — Etlworks runs in your own AWS, Azure, or GCP environment.
Best fit when you have strict compliance, data residency, network isolation, or infrastructure standardization requirements.
Hybrid (Cloud + Integration Agent)
Etlworks runs in the cloud, while a lightweight Integration Agent is installed inside your network to securely connect on-premise data sources (databases, file shares, internal APIs) to the cloud platform.
Best fit when you want managed cloud convenience but need to integrate data sources that can't be exposed publicly.
Managed Kubernetes Cluster
You own the Kubernetes cluster — self-managed, Amazon EKS, Azure AKS, or Google GKE — and the Etlworks installer manages the deployment: automated delivery, in-app scaling and pod sizing, application upgrades, offline backup and recovery, and migration between regular and managed deployments. PostgreSQL and Redis are provisioned by the installer, or brought as your own external services.
Best fit when Kubernetes is your infrastructure standard and you want elastic capacity with day-two operations in the product instead of runbooks. A 14-day Kubernetes trial is available from the downloads page. See Install Etlworks on Kubernetes.
Advanced Deployment Options
High Availability (HA)
- Supported for both on-premise and dedicated cloud deployments.
- Uses multiple Etlworks nodes plus shared infrastructure (database, Redis, shared filesystem or object store).
- Each node requires a separate license.
HA is the right choice when you have uptime SLAs, fault-tolerance requirements, or run mission-critical pipelines that can't tolerate node-level downtime.
Multi-Environment (Dev, QA, Prod)
- Etlworks supports running multiple instances for development, testing, and production.
- Environments can be deployed and managed independently or centrally.
- Flows and configurations can be moved between environments using built-in export/import — or managed as code in Git and promoted with plan-and-apply deployments (Data Integration as Code).
Multi-environment setups are typical for enterprise teams with a formal release process or strict change management.
Questions to Help You Decide
- Where should Etlworks run — in your environment or ours?
- Does Etlworks need access to internal systems (databases, file shares, internal APIs)?
- Do you have uptime or fault-tolerance requirements that call for HA?
- Do you already run Kubernetes — and should Etlworks run there as a managed cluster?
- Do you need separate environments for development, QA, and production?
- Who will manage the infrastructure — your team, Etlworks, or an implementation partner?
How Implementation Typically Works
- Scoping — A short conversation to understand your environment, integration requirements, and constraints.
- Deployment selection — Decide on the best-fit model (cloud, on-prem, hybrid) and any advanced options like HA or multi-environment.
- Provisioning or installation — Either we provision your cloud instance, or you install Etlworks in your environment with our guidance.
- Initial configuration — Set up connectivity to source and target systems, configure authentication, and validate the environment.
- First flows — Build and test the first integration flows, often using prebuilt templates as a starting point.
- Training and handoff — Once the environment is stable, your team takes ownership of ongoing development and operations.
Most cloud deployments are live within hours. On-premise and hybrid deployments typically take a few days, depending on infrastructure access and security review timelines.
Resources
Not Sure Which to Choose?
If you'd like help selecting the right deployment model for your situation, email sales@etlworks.com with a brief description of your environment and requirements. Implementation partners can also help with end-to-end deployment and configuration.