Case Study: Removing Long-Term Retention Cost from a Legacy Workload

Problem

A council organisation had a valid requirement to keep some legacy workload data for up to seven years. The business need was not the problem. The cost of keeping the environment alive was.

The workload had low active use, but it still carried production-style Azure cost. Storage, security, networking, logging, and retained platform components were all adding to the monthly bill.

Context

The organisation did not have an internal FinOps function, so there was no clear owner regularly challenging whether the retained environment still matched the business requirement.

The customer did not have clarity around cloud costs. The internal engineers had performed fundamental platform-driven cost optimisation activities and vendor engagements, but the estate still carried patterns common in large enterprises that migrated on-premises workloads into cloud without taking advantage of the variable cloud model, treating it like another hosting data centre.

Initial cost data showed storage as the obvious target, but the review had to go deeper. The environment also generated millions, and in some areas tens of millions, of storage operations. That meant transaction cost, monitoring, Defender coverage, and supporting platform design all needed to be checked before recommending changes.

Among other wins, we discovered a traditional VM worker hammering the cloud file system, constantly walking through folders, checking files, and reading metadata.

The customer was effectively paying several times for the same inefficient pattern: once for the storage activity itself, then again for security monitoring, logging, and related cloud operations.

Because the cost was spread across several services, it had blended into the normal cloud baseline years earlier.

What was accomplished

Our team reviewed the Azure cost data, storage patterns, operational requirements, and the technical design around the retained workload.

Key decisions and trade-offs

The important decision was not to remove retention. It was to stop paying premium running cost for data that mostly needed to be retained, not actively served as a normal production workload.

From there, we identified a more suitable technical approach using existing technology already available to the customer. The aim was obviously not to remove the process. But to deliver the same outcome without generating unnecessary cloud operations and cost.

We also avoided treating the first saving as the only opportunity. The review identified a further design path with another six-figure annual saving opportunity, but that required a separate business and technical decision.

Result

The agreed changes reduced the cost of the retained workload by around 90% while keeping the retention requirement intact.

The work also gave the organisation a clearer way to discuss cloud cost. Instead of a broad monthly bill, the team could see which parts of the legacy design still had value, which parts were only historical, and which parts needed a more deliberate decision.

Conclusion

Long-term retention does not have to mean long-term premium storage cost.

If your Azure environment carries old workloads, retained data, or unclear storage growth, an azure cost optimisation audit can usually find practical options without turning the review into a large programme.