Helpful Fivetran Tutorials
September 28, 2026 · SyncSpur Team
MAR is only half your Fivetran bill. The other half is the Snowflake compute the loader burns on COPY and MERGE every sync. How to attribute credits per connection, pick sync tiers, size the warehouse (usually down), make merges prune properly, and put a guardrail on it all.
Read moreSeptember 21, 2026 · SyncSpur Team
Analysts keep asking where a column came from and nobody can answer without opening the Fivetran UI. A practical guide to Fivetran's Metadata API and Platform Connector lineage tables: what metadata you actually get, how to push it into Atlan, Collibra, Alation, OpenMetadata, Unity Catalog or Snowflake Horizon, and how to stitch source-to-dbt-to-BI lineage without hand-maintaining a spreadsheet.
Read moreSeptember 21, 2026 · SyncSpur Team
The first sync is the one that hurts - days of runtime, a MAR spike that lands in month one, and a source database the DBA wants unplugged. A practical playbook for planning historical backfills in Fivetran: priority-first sync, replica selection, table and column pruning, chunking large tables, watching for re-sync triggers, and knowing when a one-off bulk load beats the connector.
Read moreSeptember 21, 2026 · SyncSpur Team
Vendor drops, partner extracts and finance exports still arrive as files. A practical guide to Fivetran's file connectors - S3, Azure Blob, GCS, SFTP and Google Drive - covering folder and file pattern design, CSV vs Parquet schema inference, compression, late-arriving and re-delivered files, upsert keys, and what all of it costs you in MAR.
Read moreSeptember 18, 2026 · SyncSpur Team
Document databases do not have a schema and your warehouse insists on one. A practical tutorial on Fivetran MongoDB syncs: change streams vs collection scans, sizing the oplog, choosing packed vs unpacked schema inference, why embedded arrays dominate your MAR bill, and the dbt staging pattern that turns _fivetran_values into columns analysts can join.
Read more