How to Calculate Sync Speed
Use Case
You want to know the estimated time it takes to sync data from your source to your destination. This requires the use of our free Fivetran Platform Connector and the sample query we provide.
Fivetran Recommendation
We recommend using the results of the following calculation method as an estimate, because sync speed depends on a variety of factors:
- Volume of data in the data source
- Connector type (for example, application, database, function)
- Subscription tier (for application connectors)
Use our Check Daily Modified Records sample query to determine the amount of records modified in a certain amount of time.
NOTE: You can change the time period used for the calculation from days to hours to minutes, depending on the volume of data.
Do the following to roughly estimate the sync speed:
Make sure that the Fivetran Platform Connector is configured and active in your account.
Run an initial sync or historical re-sync for the connector whose sync speed you want to know.
i. (Optional) Start with a few select tables. If your data source contains a large volume of data, we recommend starting with a small sample base.
Use the Fivetran Platform Connector and the Check Daily Modified Records sample query provided above to measure the average number of rows updated ('records_modified') per 24 hours.
Estimate the expected source volume - either in total records or daily records.
To project how long your Fivetran syncs will take, divide your estimated number of records (daily or total) by the average number of rows figure you found in step 3.
Considerations
Database connectors have throughput limits that can be found in our documentation for the respective database type.
If you are concerned about MAR, create a designated test connector and use it to run these speed calculation tests before creating a production connector that will sync all of your data.