Fivetran

Use it when

• Pipeline maintenance should be the vendor's problem: connectors, schema drift, and API changes are handled for you.
• Sources span SaaS applications, databases, files, and events, and the destination is a warehouse or a managed Iceberg or Delta lake.
• Database replication needs log-based CDC without operating CDC infrastructure.
• The team is small relative to the number of sources it must integrate.

Think twice when

• Monthly-active-rows pricing meets high-churn tables; cost scales with change volume and is hard to predict up front.
• You want to self-host or modify connectors; there is no self-hosted option, unlike Airbyte or Meltano.
• Sub-minute freshness is required; the fastest sync floor is one minute, on Enterprise.
• A long tail of niche sources matters; custom connectors run as Python in Fivetran's cloud rather than anywhere you choose.

How it runs

There is nothing to deploy: Fivetran runs the pipelines, with hybrid deployment for data that must stay in your network. Syncs are scheduled in the platform from one minute to 24 hours, configured in the UI or through the REST API and Terraform provider. Transformations are Quickstart models or dbt Core runs after load.