You've created your project, added a data source but it isn't importing. Now what? We've listed all the problems we've seen to date, with common workarounds and solutions for them.

If this guide does not resolve your issue, please reach out to us and we'll jump on a call.

To start with, what kind of integration isn't working?


If your connector is listed below, we recommend you start there for the most common issues:

If this doesn't help, or your connector isn't listed, here are our more general troubleshooting points.

Authorisation page errors

When you choose a connector, you are re-directed to an authorisation page hosted by Fivetran. Each connector has a unique flow to authorise your account, choose data and possibly other configuration. The most common issues tend to happen on this page, e.g. a complex error message.

While working with Fivetran uniquely enables us to give you access 150+ connectors enterprise grade connectors, the downside is that the UI is ... let's say janky and sometimes inscrutable 😅. (They wouldn't disagree with me and we're frequently giving feedback on how to improve that.) In some cases, it is a simple case of turning if off and on again (i.e. retrying) and it will work, so always try that.

If the error persists, it usually means something real is going on but it can be hard to work out what it is. If it's not obvious from the message, reach out to us and we'll jump on a call. For example, with Google Ads you need to have a manager account. It's simple once you know it, but it isn't (in our opinion) well explained.

If you're the kind of person who likes to debug themselves, take a look at the Fivetran docs (e.g. Google Ads) for your connector. The answer is usually in there.

Finally, it is possible there is a bug at Fivetran's end. We're only ever seen this once, but it's worth noting.

The good news is once you've completed the initial setup, that data will be flowing in reliably multiple times a day and you'll never have to think about it again.

Load page errors

After the authorisation page has completed, setup tests will run. Typically these will take a few seconds to five minutes, depending on the source.

The setup tests are designed to validate that the connector will work. If it fails after this during the import process, something has gone wrong and you should reach out to us.

Google Analytics

Google Analytics 4 is not currently available, as their new reporting APIs are unstable. This is a popular request and we'll share an update as soon as it's ready.

Google Ads

For the Google Ads connector, your account needs to be within a Google Ads Manager account hierarchy. This is required even if you only have one account, as only Manager accounts have access to the API to extract data. The Fivetran authorisation page does not explain this well and there are no error messages if you forget to do it, it just fails to show any accounts.

You can read detailed instructions on how to do that here.


The Postgres connector will check for a valid SSL certificate. In some situations, you will get a pop that says "Confirm TLS Certificate", and the option to select a trusted certificate. When you press confirm, the validation process will restart - this is normal.


For the Twitter connector, the authorisation page will use the currently selected Twitter account in your browser. You might need to switch account in the same browser as you're using Gyana prior to authorisation.

Google Sheets

All rows are blank

If you don't add an explicit cell range, the Google Sheets integration will try to infer the cell range in the first sheet by looking at cells with content.

It's possible that if your Google Sheet has a blank cell or a cell with whitespace, this will be included and distort the cell range. For example, we saw an example where a 100 row dataset had a blank cell at row 10,000, and so the imported data had 9,900 blank rows.

The solution is either to copy the data into a new sheet (and delete the old sheet), or explicitly provide a cell range.

The header is wrong

For the Google Sheets integration, the first row of the data must be the header values (i.e. the column names). If you don't include this, the first row of data will become the column names.

Schema auto-detection fails

When we import a Google Sheet, we use the top 500 rows of data to infer the schema. If for example, the top 500 values for a column are a number, but further down there is a string, you might get an error like Could not convert value to integer. Row XYZ, Col AB.

In the short term, the solution is to manually move that row to the top and repeat the import. We're work on a fix to avoid this error in future.

If you think you've found any other issues with Google Sheets, reach out to us.

Upload CSV

The CSV file format is not fully standardised, which means there is a huge variety in the formatting of data outputted from different systems (e.g. how are quotes escaped, how are newline characters treated).

Our CSV importer is designed to be permissive as possible, e.g. accepting jagged rows and quoted newlines.

If your CSV fails to import, here are a few steps you can try:

  • Make sure your CSV file has a header row. This is required for our importer. (If not, open with Excel/Google Sheets and add one.)

  • Read the error message and see if it helps.

  • Check your CSV is actually "comma-delimited". If the separator between records is a tab or pipe (|), you can set that in the Configure advanced settings.

  • Upload the CSV file to Google Drive, open as a Google Sheet and use the Google Sheets integration. (This works because Google Sheets is particularly good for problematic CSV files.)

If none of these steps works, reach out to us. We'd also love to hear from you even if the Google Sheets workaround worked, to see if we can improve the CSV upload experience in future.

Did this answer your question?