Blob store performance

[TJbrunk]

Summary

Improves blob store performance and observability by pre-filtering CID input lists, enhancing error logging diagnostics, and making batch operation concurrency configurable.

Changes

Pre-filtering & Deduplication (mod.rs)

• Filter out common non-CID strings (urn:uuid:, did:key:) before batch operations
• Remove duplicate CIDs to reduce redundant Azure Blob Storage requests
• Add trace log of the final CID list being fetched

Configurable Batch Concurrency (mod.rs, blob_store.rs, azure_blob_bulk_get.rs)

• Add optional concurrency: Option parameter to:
  • exists_many()
  • get_many()
  • put_many()
• When None, falls back to batch_concurrency_limit() (default: 16)
• Updated all call sites (FFI layer, example, tests) to pass None to use defaults

Enhanced Error Logging (azure_blob.rs)

• Distinguish 404 responses (blob not found) from other HTTP errors with appropriate log levels
  • 404 → debug-level log (expected case)
  • Other HTTP errors → warn-level with status code and error_code details
  • Non-HTTP errors → warn-level with error kind details
• Fix typo in error message ("not init" → "not initialized")
• Provide structured logging with CID and error details for easier debugging

Impact

• Reduces unnecessary API calls to Azure Blob Storage by filtering invalid inputs and deduplicating requests
• Enables fine-grained concurrency control per operation while maintaining backward compatibility
• Improves troubleshooting with more granular error classification and detailed context