Blob Expired but Still in Storage? Here’s the Cleanup Flow I Built

When I was building SnappShare, a file-sharing tool with auto-expiry, I ran into a subtle but critical issue:
Files were expiring for users, but not leaving storage.

Let me break down what happened, why it mattered, and how I fixed it using Azure Service Bus and Azure Functions.

---

The Problem: Lifecycle Wasn’t Precise Enough

Azure Blob Storage supports lifecycle management policies, but here’s the catch:

• The minimum delay for deletion is 24 hours.

• In SnappShare, users could share files that expired in 1 hour (using short-lived SAS tokens).

• While access was blocked, the actual file remained in storage.

This isn’t an issue at low scale. But at high volume, you could get charged for redundant files that users can no longer access.

---

The Goal: Precise, On-Time Deletion

I wanted a system that:

• Deletes files at their actual expiry time

• Scales well and doesn't run 24/7

• Offers visibility and error handling

---

The Solution: Custom Cleanup Flow

Here’s the simple but powerful workflow I implemented:

1. When a file is uploaded, I store the expiry time (e.g., 2 hours from now).

2. I send a scheduled message to an Azure Service Bus Queue that will unlock at that expiry time.

3. When the message is unlocked, an Azure Function is triggered to:

   • Delete the blob

   • Log success or failure

4. If needed, I can plug in a Topic + Subscription system to send expiry notifications.

---

Why This Works Better

• Precise timing: Files are deleted exactly when they expire

• Pay-as-you-go: Service Bus (Basic) + Azure Function (Consumption) = cost-efficient

• Built-in retries: Functions can retry failed deletes

• Observable: Add logs, alerts, or even notify users when cleanup happens

---

Is It Overkill?

You might ask: Why not just let lifecycle do its job?

Here’s why:

• You don’t always want to wait 24+ hours

• You don’t want surprises in your storage bill

• You want control

Designing for precision helps keep your product sharp, fast, and scalable, especially for file-heavy SaaS.

---

Final Thoughts

Azure Blob Storage is powerful, but when you need more control over cleanup, pairing it with Service Bus and Functions gives you the power to design around failure and scale with confidence.

Built something similar or curious how it works under the hood?
Let’s talk on X → @freemancodz | Full repo → GitHub

#azure #blobstorage #serverless #servicebus #dotnetdeveloper #snappshare #cleanarchitecture #cloudengineering