Azure Blob Storage destination: common problems and their solutions

This guide outlines users’ most common issues when integrating Windsor.ai data into Azure Blob Storage. It covers authentication errors, data mismatches, upload failures, scheduling problems, and formatting issues, and shares troubleshooting steps and best practices.

Issues:

• “Test Connection” fails in Windsor.ai
• Destination task remains in “failed” or “not started” state
• Azure Blob container is not created, or data is not streamed
  

Possible causes:

• Invalid or expired connection string
• Incorrect container name
• Azure storage account is inactive or deleted
• Incomplete setup of the resource group or storage account
  

Solutions:

• Double-check the connection string:
  • Go to Azure Portal -> Storage Account -> Access keys.
  • Copy the full connection string from Key 1 or Key 2.
  • Paste it exactly as-is into Windsor.ai.

• Verify container name:
  • Ensure the container exists in your storage account.
  • If using a new name, Windsor.ai can auto-create it.
  • If the name is invalid (e.g., with uppercase letters or underscores), Azure will reject it.
• Check Azure subscription and resource availability:
  • Confirm that your Azure subscription is active.
  • Ensure the storage account and container have not been deleted.
• Retry authentication:
  • If errors persist, delete the destination task in Windsor.ai and recreate it with correct credentials.

Issues:

• Files do not appear in Blob Storage
• Upload task status is OK, but no data is visible
• Wrong folder or blob naming convention
  

Possible causes:

• Invalid folder path format
• Inconsistent blob prefix
• Expecting data in the wrong folder
  

Solution:

• Check folder path format:
  • Folder paths should not begin or end with a slash (e.g., marketing/ga4, not /marketing/).

• Validate blob prefix:
  • Blob name is used as a prefix in the file name.
  • Ensure this is unique per data source or consistent across your team.

• Check in Azure Portal:
  • Go to your container and check if files are listed under the correct path and name.

Issues:

• Data uploads only once
• No updates after the first successful run
  

Possible causes:

• Schedule type not saved in the destination task
• Plan limits exceeded
• Internal errors prevent retries
• Cache issues
  

Solutions:

• Set a schedule explicitly:
  • When creating the destination task in Windsor.ai, choose a Daily or Hourly schedule type and click Save.

• Check your Windsor.ai plan tier:
  • Free/Basic: Daily updates only
  • Plus/Professional: Hourly updates available
• Monitor run status:
  • In your Windsor.ai’s Azure Blob Storage destination, check the affected Destination Task -> Logs
  • Confirm if the schedule was triggered and completed by clicking on the “!” icon beside the task, and see error logs in the destination task form.

• Clear Windsor.ai cache:
  Windsor caches data for 6 hours. Click “Clear cache” in the data preview to fetch fresh data if recent changes aren’t visible in the blob storage.

Issues:

• Files overwrite previous files
• Data appears in unexpected locations
• Duplicate uploads or broken folder structures
  

Possible causes:

• The same blob prefix used across multiple data destination tasks
• Folder path misconfigured
• Slash characters used improperly
  

Solutions:

• Use unique blob name prefixes for every destination task:
  • Example: ga4_analytics_, fb_ads_, etc.
• Correctly format folder paths:
  • Use a format like data/google-analytics or marketing/2024/
  • Avoid leading/trailing slashes.
• Avoid reusing task configuration:
  • When cloning or editing tasks, make sure to update the blob prefix and folder path accordingly.

Issues:

• Downloaded files don’t open correctly (e.g., broken CSV)
• Missing columns or malformed data
  

Possible causes:

• Field selection issues in Windsor.ai
  
• Source schema changed after the task setup
• Export format mismatch
  

Solutions:

• Preview data before exporting:
  • In Windsor.ai, go to Data Preview and check that all necessary fields are selected.
• Recreate destination task:
  • If fields were added or renamed in the source, delete and recreate the Azure Blob destination task.
    
• Validate CSV file:
  • Open it in a text editor to inspect the structure.
  • Use Azure Storage Explorer or download from the Portal.

Issues:

• Metrics not matching the source platform
• Unexpected data patterns or gaps
  

Possible causes:

• Wrong filters or data source accounts selected
• Wrong date range selected
• Manual schema override in Azure
  

Solutions:

• Use consistent filters and data sources:
  • Ensure Windsor.ai’s selected accounts/data sources/date ranges, and filters match what you expect.

• Compare with raw API data:
  • Use Windsor.ai’s API to fetch the same data manually and validate.

• Windsor.ai Task Logs:
  • Check task run status, errors, and schedule execution.

• Azure Storage Explorer:
  • View container contents, validate file structure, and naming.

• Azure Portal Activity Logs:
  • View authentication, upload, or permission errors in a storage account.
  • Go to your storage account and under “Monitoring” select “Insights” and see if there are any failures/errors reported.

• Windsor.ai Data Explorer:
  • Manually test data queries to compare outputs and preview in the data preview dashboard.

By following these troubleshooting steps, users can resolve common issues when streaming data to Azure Blob Storage via Windsor.ai.

If the issue persists, contact our customer support for further assistance.

Keep in mind that investigating data discrepancies takes time (1-3 days), so it’s better to organize the report and send it to [email protected]. We’ll take a detailed look and reply to you as soon as we find a resolution.