We recently automated our invoice archival process using Azure Logic Apps and Blob Storage, eliminating a manual workflow that consumed significant team hours. Previously, our finance team manually exported invoices from our SQL database, renamed files with metadata tags, and uploaded them to storage every week.
The solution uses Logic Apps to trigger daily at 2 AM, query completed invoices from the past 24 hours, extract invoice metadata (customer ID, invoice date, amount), and upload PDFs to Blob Storage with systematic naming: invoices/{year}/{month}/{customerId}_{invoiceId}.pdf. We implemented lifecycle management policies to automatically move invoices older than 90 days to Cool tier and archive after 7 years for compliance.
Key benefits: 80% effort reduction (from 6 hours weekly to automated), zero human errors in file naming, consistent metadata tagging, and automatic cost optimization through tier transitions. The entire workflow processes 200-300 invoices daily without intervention.
Excellent implementation! The lifecycle management integration is particularly smart for cost optimization. Are you using blob index tags or metadata properties for the customer and invoice attributes? Index tags would enable faster queries if you need to retrieve specific invoices later without scanning the entire container hierarchy.
For archival-only scenarios, metadata properties are perfectly adequate and what you’re doing is correct. Blob index tags shine when you need to query across containers or perform complex filtering without knowing exact paths. Since your ERP maintains the searchable records and blobs are compliance archives, metadata is the right choice-simpler and no additional indexing overhead.
We’re using blob metadata properties currently. The Logic Apps workflow sets three metadata fields during upload:
CustomerId: {value}
InvoiceDate: {ISO8601}
Amount: {decimal}
We haven’t needed fast querying yet since our ERP system maintains the searchable database. The blob storage is purely archival for compliance and backup. Would index tags offer significant advantages for our use case, or is metadata sufficient for archival scenarios?