Latest / The Cloud Business Podcast with Fexingo: AWS, Azure, GCP, and Enterprise Infrastructure / How Cloud Bills Now Meter Your Data Backup Storage
Transcript
- Lucas: So Luna, I was looking at a mid-sized SaaS company's cloud invoice last week, and the line item that jumped out at me wasn't compute or egress. It was backup storage. Luna: Interesting. That usually gets lumped into 'storage' without much thought. What did you see? Lucas: Their backup bill had grown about thirty percent quarter-over-quarter, and when I dug in, the charge wasn't for raw gigabytes. It was for snapshot versions. Every time they took a daily snapshot, the delta was being billed, and they were keeping ninety days of them. Luna: So it's not just about how much data you have, but how often you copy it and how long you keep it. That's a different mental model. Lucas: Exactly. And this is the shift we're seeing across the big three — AWS, Azure, Google Cloud. They've moved from simple per-gigabyte pricing to metering the lifecycle of each backup. You're charged for the initial snapshot, then for every incremental change, and increasingly for every time that data is read back during a verification or a restore. Luna: So the bill is now tied to how often you poke at your backups, not just what you store. That's a subtle but important change. Lucas: Right. Let's take AWS as an example. AWS Backup charges based on the storage used, but also for each restore and for cross-region copies. But what caught this company off guard was the cost of snapshot deltas in Elastic Block Store, or EBS. Each snapshot only stores the changed blocks, but if you're taking daily snapshots on a database that churns constantly, those deltas add up fast. Luna: And I bet the invoice just shows 'EBS Snapshots' with a big number. You have to dig into the console to see the per-snapshot breakdown. Lucas: That's the problem. The pricing page shows per gigabyte per month, which sounds cheap, but the metering is per snapshot, per retention period, and per restore event. The CFO sees a line item that's grown forty percent year-over-year, but there's no single 'backup usage' dashboard that ties it to actual data changes. Luna: So how does this play out in Azure and Google Cloud? Are they following the same playbook? Lucas: Azure Backup has a similar model — you pay for the protected instances, the storage consumed, and now they've started charging for 'backup integrity checks' that run automatically. Google Cloud's Backup and DR service bills for the actual storage used, but they also charge for each backup job and for restores. What's notable is that all three are introducing more granular metering around backup verification. Luna: Backup verification reads — that's the new battleground. I've seen that term on a few invoices this year. Lucas: Yeah. These are the background scans that ensure your backups are restorable. They read the data back, check checksums, and simulate restores. Historically that was bundled into the storage price, but now it's a separate metered line item. For a company with, say, ten terabytes of backups, those verification reads could add a few hundred dollars a month. It's not huge, but it compounds. Luna: It's like the cloud providers are turning every internal process into a billable unit. But I get it — they're investing in faster restore times and better encryption, so someone has to pay. Lucas: And that's the counterargument. These fees fund tangible improvements. Azure now guarantees a certain restore speed tier, and Google's backup service offers tamper-proof storage with immutability — that comes with a premium. The question is whether enterprises are getting value for the new charges, or just paying for the privilege of using the API. Luna: So what's the practical takeaway for someone who's about to pay this bill? How do they keep from getting blindsided? Lucas: First, audit your backup settings. That company I mentioned had enabled daily snapshots on all their RDS databases without checking the default retention. They were keeping ninety days when their recovery point objective only required thirty-five. Halving that retention cut their backup bill by a third. Luna: That's a quick win. But what about the incremental changes — how do you forecast those? Lucas: You have to look at your data churn rate. If your database changes ten percent a day, your snapshots grow by ten percent daily, and that's your true cost driver. Tools like AWS Cost Explorer can show you snapshot storage per resource, but you have to map that to your change rate. That's something finance teams rarely do. Luna: So it's about understanding your own data, not just the price list. And maybe moving cold backups to archive tiers? Lucas: Exactly. For backups older than thirty days, you can transition them to cold storage like S3 Glacier or Azure Archive. The retrieval cost is higher, but if you're not restoring them often, it's a massive saving. But here's the catch — you need to model the retrieval time. If you have a compliance requirement to restore within hours, cold storage won't work. So it's a trade-off between cost and speed. Luna: That's the kind of nuance that gets lost in the 'just move to cold storage' advice. You have to align your retention with your actual recovery needs. Lucas: Right. And the other lever is negotiating with the provider. Backup storage is often a separate line item in enterprise agreements, and it's negotiable if you have volume. One client of mine got a twenty percent discount on backup storage by committing to a three-year term. It's not as visible as compute, but it's on the table. Luna: I've also heard of companies using 'backup as a service' from third-party vendors that might have cheaper egress or no verification fees. Is that worth exploring? Lucas: It can be, but you have to weigh the integration complexity. Third-party tools like Veeam or Commvault are great for multi-cloud environments, but they can charge their own licensing fees. The key is to do a total cost of ownership analysis — not just the storage cost, but the cost of managing it, the restore time, and the risk. Sometimes the native tool is cheaper even with the new metering. Luna: And there's always the human element — someone has to understand what these fees mean and adjust behavior. That's a good segue to something I've been thinking about. Lucas: What's that? Luna: Well, this kind of insight — knowing how to audit your cloud bill and spot these hidden charges — is exactly what our listeners rely on. And if they find these conversations useful for what they're building or running, supporting the show is a great way to keep it going. Lucas: Yeah, that's true. We're independent and ad-free, and that's only possible because of the listeners who chip in. If you've gotten value from episodes like this, consider buying us a coffee at buy me a coffee dot com slash fexingo. It's a small gesture that goes a long way. Luna: And it helps us keep diving into these nooks and crannies of cloud pricing. So, back to backup storage — one more angle I want to get your take on. Lucas: Sure. What's on your mind? Luna: You mentioned backup verification reads. How do those actually work? Is it like a test restore that runs in the background, and if so, does that count as a restore on the invoice? Lucas: Great question. In most cases, verification is a metadata check or a partial read, not a full restore. AWS, for instance, uses automated checks that sample data blocks and compute checksums. But they bill per GB read during those checks. So a company with a lot of backups might see a separate 'Backup Verification' line item, which was previously hidden. Luna: So the cost scales with how often you run those checks, which is configurable. Some providers let you set a verification schedule — daily, weekly, monthly. Lucas: Right. And the default is usually daily, which is overkill for most workloads. If you drop it to weekly, you can cut that cost by eighty percent. That's a low-effort optimization that most companies miss. Luna: So the real theme here is that cloud billing is getting more granular, and you have to be intentional about every setting. It's no longer just 'set it and forget it'. Lucas: Absolutely. And the good news is that the tools are getting better. AWS has a backup usage report, Azure has cost analysis for backup, and Google has a billing export that breaks down each operation. The challenge is that they're not connected — you have to manually pull the data and map it to your business metrics. Luna: Which is exactly where a FinOps practice comes in. Someone needs to own this monthly, not just during contract renewals. Lucas: Yeah. And if you don't have a dedicated FinOps person, at least set a quarterly reminder to review your backup settings. Changing retention from ninety days to thirty-five, dropping verification from daily to weekly, and moving old snapshots to cold storage could easily cut your backup bill in half. Luna: That's a solid takeaway. I know I'm going to check our own backup setup after this. Lucas: And that's the whole point. These charges aren't malicious — they're just new, and they need to be actively managed. The companies that treat backup as a strategic cost rather than an afterthought will be the ones who don't get a surprise invoice next quarter. Luna: Well said. And for anyone listening, if you want to see a breakdown of how these fees work across providers, we'll have the show notes up with links to the pricing pages. Lucas: Yeah, we'll include a checklist for auditing your own backup bill. Until next time, keep an eye on those snapshots.