Export all Hevo pipelines, sources, destinations, and configurations. For each pipeline, document: source type, destination, tables/columns synced, sync frequency, and any transformations. Use Hevo's UI to capture this. Create a spreadsheet with columns for each attribute. Note any custom webhooks or API integrations.
⚠️ Watch Out For:
- Hevo's transformation features (if used) don't map directly to Airbyte—identify these early
- Custom webhooks and scripting logic may require separate implementation in Airbyte
Decide: Airbyte Cloud (managed) or self-hosted (Docker, Kubernetes). For cost savings, self-hosted is compelling. For operational simplicity, Airbyte Cloud is recommended. Document your choice and infrastructure assumptions.
⚠️ Watch Out For:
- Self-hosted Airbyte requires Docker/Kubernetes knowledge—budget accordingly if new
- Airbyte Cloud has rate limits—verify they match your pipeline volume
Set up Airbyte (Cloud: sign up; self-hosted: docker run or kubectl). Create workspace. Configure destination connections for all target warehouses (Snowflake, BigQuery, Redshift, etc.). Test warehouse connectivity.
⚠️ Watch Out For:
- Self-hosted setup takes longer than Cloud—allow extra time for PostgreSQL, networking, and deployment
- Warehouse credentials must have schema/table creation permissions
Select the simplest Hevo pipeline (one source, one target). Create the equivalent Airbyte source connector. Configure table/stream selection and column filtering to match Hevo. Set sync mode (full vs. incremental). Run a test sync. Compare outputs with Hevo.
⚠️ Watch Out For:
- Airbyte connector versions vary in maturity—use stable, certified connectors
- Some sources require additional configuration (cursors, API keys)—read documentation carefully
If your Hevo pipelines include transformations, replicate them in dbt. Create dbt models that build on Airbyte-loaded tables. Configure dbt to run after Airbyte syncs (via webhook or orchestrator). Test dbt output matches Hevo's transformation results.
⚠️ Watch Out For:
- Hevo's built-in transformation logic doesn't map directly to dbt—requires SQL rewriting
- Scheduling dependencies between Airbyte and dbt must be carefully managed
Set up Airbyte sync schedules to match Hevo's frequency. Configure alerts (Slack, email, webhooks) for failed syncs. Monitor sync success rates and latency in Airbyte UI. Set up dashboards to track pipeline health.
⚠️ Watch Out For:
- Airbyte's schedule granularity may differ from Hevo's—adjust expectations
- Alert tuning is important—avoid alert fatigue while catching real issues
Progressively migrate remaining Hevo pipelines to Airbyte. Start with simple ones. For each, validate that outputs match the original Hevo pipeline. Document any pipelines that couldn't be directly migrated and why.
⚠️ Watch Out For:
- Most pipelines should be straightforward—if you hit complexity, investigate whether the pipeline logic was more sophisticated than expected
- Some Hevo sources may not have Airbyte equivalents—plan alternatives
Keep both Hevo and Airbyte pipelines running in parallel for 1-2 full sync cycles. Compare record counts, data accuracy, and timing. Validate that dashboards and reports produce identical results with Airbyte data.
⚠️ Watch Out For:
- Timing mismatches between Hevo and Airbyte can complicate comparison—align schedules temporarily
- Small data discrepancies (NULL handling, data type precision) often surface—investigate thoroughly
For Airbyte Cloud: review costs (per connector, per sync). For self-hosted: monitor infrastructure costs (compute, storage). Optimize Airbyte connector configurations (column selection, sync frequency) to manage costs. Update team documentation.
⚠️ Watch Out For:
- Self-hosted Airbyte has hidden infrastructure costs (servers, databases, bandwidth)—account for these
- Airbyte Cloud costs may be comparable to Hevo depending on volume—carefully evaluate the business case
Once Airbyte passes validation, disable Hevo pipelines. Keep Hevo running read-only for 1-2 weeks for reference. Update team documentation and runbooks. Cancel Hevo subscription or reduce to minimal tier. Archive Hevo pipeline configurations.
⚠️ Watch Out For:
- Don't cancel Hevo subscription immediately—keep it running for 1-2 weeks in case rollback is needed
- Update team onboarding documentation to reference Airbyte instead of Hevo
