The Curious Case of the Exploding Snapshot Costs

That's the question my manager sprung on me last quarter. Seemed like a simple enough request, but it turned into a deep dive into the quirks of our deployment process and the hidden costs of "playing it safe."

Our team, like many others, had a robust disaster recovery plan. Too robust, as it turned out.

Redundancy Gone Wrong

Before each deployment, we created multiple snapshots of our managed disks. The idea was that if anything went wrong, we could roll back to any previous state. In theory, this was sound. In practice, we rarely, if ever, used anything beyond the most recent snapshot.

We had snapshots upon snapshots, piling up like digital dust bunnies in our Azure subscription. Each deployment generated at least three new snapshots per disk, and we deployed frequently.

Incremental Backups to the Rescue

EazyOps flagged not only the redundancy but also the opportunity: incremental backups. Instead of creating full snapshots each time, we could just back up the changes since the last snapshot. This significantly reduced the storage footprint and, more importantly, the associated costs.

Lessons Learned

• Redundancy isn't always resilient: Blindly creating multiple backups can lead to unnecessary costs.
• Visibility is key: You can't optimize what you can't see. Tools like EazyOps provide the insights needed to identify cost-saving opportunities.
• Small changes can have a big impact: Switching to incremental backups was a relatively simple change with significant financial benefits.