You’re already paying for it. You just aren’t using it.
Part 1: Why your backup environment is quietly throwing money away
You pay your Azure invoice dutifully every month. The backups run. Green checkmarks in the console. And yet, in nearly every organization OptimaSure audits, Edco Wallet sees the same pattern: the backup environment keeps growing, quietly, unmanaged, and unseen. And the costs grow right along with it. In this first part of a two-part series on Azure Backup, he looks at the technical cost pitfalls.
What strikes me: most organizations set their backups up properly at some point. At the moment of initial configuration, everything checked out. The problem is that nobody looks at it again after that. New workloads get added, old ones disappear, but the backup configuration never moves with them. Meanwhile, the costs keep creeping up, month after month, without anyone noticing.
Most IT managers know that storage in Azure costs money. What’s less well understood is exactly how Azure Backup’s cost structure is put together.
Azure Backup runs primarily on Recovery Services Vaults. You pay per protected instance, and for the data stored. Sounds straightforward. In reality, there are four separate cost lines: data storage for the backup data itself (local or geographically replicated), snapshot storage for instant recovery, data egress costs when restoring backups into another region, and even backup alert notifications that can carry a cost.
In a lot of organizations, those four lines never get looked at as one whole. That’s where it goes wrong. Backup gets treated as a generic service that doesn’t need any tailoring.
Azure offers several redundancy levels for backup storage: Locally Redundant Storage (LRS), Zone-Redundant Storage (ZRS), and Geo-Redundant Storage (GRS). Geographically stored backup sets cost roughly two to three times as much as locally stored backups.
The default setting in a lot of Azure configurations? GRS. Makes sense for critical production data. But when I ask whether every system under that vault actually needs geo-redundant backups, the answer is rarely a convincing “yes.”
Development environments, test servers, archive data, they’re all running on GRS, because nobody ever asked the question: “Do we actually need this?”
For an organization with 50 TB of backup data, this single choice alone can mean overpaying by €1,500 to €3,000 a month. Without anyone ever having deliberately chosen it.
Azure Backup uses instant recovery snapshots for fast restores. Those snapshots are kept alongside the regular backup data in the vault. Useful for speed, but they come with their own separate storage cost.
The problem: the retention period for those snapshots usually gets set during the initial configuration, and is never touched again after that. Two, five, sometimes seven days, when one day is plenty for most workloads.
Snapshots are, by definition, full copies, not incremental. Every extra day of retention is a full copy of the disk. For a 1 TB disk with seven days of snapshot retention, you’re paying for 7 TB of snapshot storage, every single month. That’s different from the on-prem SAN snapshots many of us know from before.
Recovery Services Vaults grow. That’s their job. But in a lot of environments, they also grow to keep pace with forgotten workloads: servers that were switched off long ago, databases nobody queries anymore but that are still technically online, whose backup configuration was never removed.
Azure charges per protected instance, even if that instance hasn’t been active in a long time. As long as the vault configuration is active, the meter keeps running.
At a recent client, we found eighteen vault items still pointing to VMs that had been deleted more than a year earlier, and 38 terabytes of databases that should have been cleaned up long ago. No active systems left — just an active cost line.
Based on our analyses at SME+ organizations with 10 to 200 VMs, we typically see 25 to 35% in unnecessary costs in the backup environment. That usually breaks down into 40-50% overkill on storage tier (GRS where LRS would do), 30-40% unnecessary snapshot retention, and 10-20% cost from forgotten or inactive instances.
For an organization spending €5,000 a month on Azure Backup, that means, on average, €1,250 to €1,750 a month that simply isn’t necessary.
The first step is insight. In Azure Cost Analysis, look at the breakdown of your backup costs at the resource level. Look specifically at which vaults are running on GRS and whether that’s actually necessary for every workload under them, what the snapshot retention is per backup policy and when that policy was last reviewed, and whether there are vault items still active for systems that no longer exist.
In part 2 of this series, we’ll dig deeper into the policy side of these costs: how an outdated retention policy and never testing your restores combine to form the most expensive pairing you can have in Azure.
Want to know what your backup environment is really costing you? Schedule an Azure Cost Scan and get concrete insight within 30 days, including an action plan.