Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
As data estates grow, organizations must manage and optimize storage across billions of objects distributed globally. Without consistent policies, unmanaged growth can lead to rising costs, operational complexity, and fragmented governance that diverts effort from higher‑value work.
Azure Blob Storage provides built-in capabilities to help you reduce storage costs by automating data tiering, retention, and large-scale management. You can use these capabilities independently or combine them to meet different operational and cost-optimization needs.
This article compares three Azure Storage capabilities for tiering and cost optimization:
- Smart tiering – Microsoft-managed adaptive tiering based on observed access patterns
- Azure Blob Storage lifecycle management – Rule-based tiering and deletion policies configured per storage account
- Azure Storage Actions – Serverless orchestration for data management tasks across multiple storage accounts
Quick decision guide
Use the following table to choose the Azure Blob Storage cost optimization capability that best fits your scenario.
| If you need… | Use |
|---|---|
| Microsoft-managed adaptive tiering when access patterns are unknown or you don't want to manage transitions | Smart tiering |
| Deterministic, age-based tiering or retention within a single storage account | Lifecycle management |
| Dynamic targeting, multi-account execution, or multiple operations in a single workflow | Azure Storage Actions |
Smart tiering: Microsoft-managed adaptive tiering
Smart tiering is a fully managed solution that automatically moves data between hot, cool, and cold access tiers based on access behavior. It minimizes manual configuration and adapts as data usage changes.
Key benefits include:
- Automatic cost optimization – Objects transition between tiers based on observed access patterns, without requiring policy rules.
- Minimal operational overhead – Well suited for scenarios where access patterns are unknown or change frequently or you do not want to manage tier transitions.
- Predictable billing model – You pay capacity at the resident tier and standard hot-tier transaction costs. There are no early deletion, rehydration, or tier transition charges for smart-tiered objects.
By default, new objects start in the hot tier. As data ages without access, it moves to cooler tiers and automatically returns to hot when accessed again. Tier-specific performance and service-level characteristics still apply.
Smart tiering includes a monitoring fee per 10,000 monitored objects per month for objects larger than 125KiB.
NOTE It may be more cost effective to keep objects smaller than 125KiB in the hot tier to avoid impact from minimum objects size billing in cooler tiers. Learn more
Azure Blob Storage lifecycle management: Rule-based tiering and retention
Azure Blob Storage lifecycle management overview provides policy-based automation for transitioning tiers or deleting blob data based on time-based conditions such as creation date, last modification, or last access.
Key benefits include:
- Deterministic rule-based automation – Define explicit policies to tier, archive, or delete data based on age or metadata.
- Low operational overhead – A single policy per storage account runs automatically without infrastructure management.
Lifecycle management is best suited for predictable retention and compliance scenarios. Policies run on a best-effort basis and don't affect application performance, but you should monitor outcomes to validate expected behavior.
Azure Storage Actions: serverless data management at scale
Azure Storage Actions provides a fully managed, no-code framework for orchestrating complex data management workflows across large storage environments.
Key benefits include:
- Conditional task composition – Apply dynamic conditions by using blob metadata, index tags, and logical operators.
- Multistep operations – Perform multiple actions on matching objects within a single task definition.
- Reusable and scalable – Apply the same task definitions across multiple storage accounts through assignments.
- Flexible execution – Run tasks on demand or on a schedule.
- Built-in monitoring – Track execution status and outcomes through dashboards and reports.
Azure Storage Actions is ideal for large-scale environments that require coordinated, cross-account data management beyond simple tiering and retention.
Example scenarios
Managing large volumes of data with varying access patterns
An enterprise accumulates terabytes of user-generated media and backup files monthly. Most files are accessed frequently for a short period, then rarely needed, but must be retained for compliance. Users must optimize costs of data storage by moving log files into appropriate access tiers. This challenge is exacerbated with larger data estates that support multiple workloads.
Azure Blob Storage lifecycle management allows you and other users to create a set of rules to move objects to cost-efficient storage tiers based on age or access. You can set up a policy to move objects to the cool tier 15 days after creation, to the archive tier 90 days after creation, or delete objects after one year without manual intervention.
Dynamic retention policies with parameterized conditions
An enterprise has a multi-petabyte data lake with ingestion containers that house raw, processed, and gold curated datasets. Only the raw partitions should age out after 30 days. The gold/ and curated/ paths must always be excluded from cleanup and tiering jobs. They need to selectively perform operations on specific blobs while ensuring flawless functionality and by avoiding bugs.
Storage Actions can help you achieve this goal. You and other users can create a parameterized rule that deletes blobs that are older than a specified number of days which is configurable in the blob tags (for example, 30, 60, or 90). You can make the task more targeted by adding rules based on the name using wildcards. Finally, applying prefix exclusions for gold/ and curated/ ensures the objects in those containers aren't processed.
The same task can have unique prefixes and schedules to run when assigned to each storage account you want to target. With one task and multiple assignments, you can dynamically adjust retention using parameters. And prefix exclusion ensures curated data is never touched, eliminating accidental deletions.
AI/ML training datasets with bursty reuse
Machine Learning (ML) platform leaders need to serve many experiments across teams. Datasets and checkpoints oscillate between "cold for weeks" and "hot today," making rule based tiering brittle and error prone. Teams either overpay in the hot tier or incur retrieval or early deletion costs when reusing data.
The best way to solve this problem is to turn on smart tier at the account scope. Datasets drift to cool or cold tiers after inactivity. When a job references a blob, promotion to the hot tier is instant and no retrieval or transition fees apply. Access transactions charge at hot transaction rates, which is ideal for training loops. No need to predict access windows or maintain per prefix rules. The engine adapts to research churn automatically and the monitoring fee is predictable.
Automated cleanup of snapshots
Development teams create frequent snapshots for testing, which can quickly accumulate and inflate storage costs.
You can configure Azure Blob Storage lifecycle management to delete all snapshots older than a set period. However, you can reduce the scope of the policy only by filtering on blob path prefixes. If the scenario needs the use of blob index tags or excludes certain path prefixes to limit the scope, then Azure Storage Actions is the ideal tool.
Regulated customer: policy-driven tagging and action pipelines
A healthcare organization must comply with regulations that require all Protected Health Information (PHI) to be securely retained for seven years. At the same time, the organization manages large volumes of non-PHI data generated by business operations, which should be efficiently tiered or deleted based on usage to optimize storage costs.
By using Storage Actions, the organization implements a data management strategy that uses tagging, filtering, and multistep action pipelines to meet both compliance and cost optimization requirements. As new data is ingested, Storage Actions tags it. All objects are assigned a default tag, such as contentType='nonPHI', unless they're explicitly identified as PHI with targeted storage task conditions.
By using a second storage task, you can define conditions such as the following:
If an object is tagged as
contentType='PHI', the engine skips tiering and applies a strict retention or immutability policy to prevent deletion, ensuring regulatory compliance.If an object is tagged as
contentType='nonPHI', a tiering operation is applied which automatically moves data to lower-cost storage.
Complex compliance and storage logic is managed declaratively, simplifying operations and reducing errors with just two storage tasks defined.
Unknown access patterns
A product team stores user-generated content, such as images, documents, and media, and the access patterns are difficult to predict. Some objects are used frequently for days, and others aren't used for months. Then, suddenly, use of those objects suddenly spikes again due to seasonal usage, investigations, or customer support needs. The team doesn't want to maintain per-prefix rules or constantly tune policies as workloads evolve.
When you turn on smart tiering at the account scope, the service automatically adapts tiers based on observed access. This feature is a good fit when access patterns are unknown or you prefer hands-off management. If objects move to cooler tiers and then become active again, smart tiering can promote them back without adding extra tier-transition, early deletion, or data retrieval charges. This feature makes the cost model simpler and more predictable.
Capability comparison
| Feature | Azure Blob Storage lifecycle management | Azure Storage Actions | Smart tiering |
|---|---|---|---|
| Best for | Simple conditions, recurring data tiering and retention needs, automating routine cost and compliance tasks. | Complex, conditional, and organization-wide data management scenarios such as applying custom governance policies, orchestrating multistep operations, or integrating with business processes without the need for custom development | Hands-off data tiering when you don’t know or don’t want to manage access patterns |
| Deployment | Policy per account, runs automatically, no control on schedule | Task with multiple assignments. Each assignment can be run on-demand or set to a specific schedule. | Account-level enablement |
| Access controls | Native storage feature | Managed identity. Admins can define one managed identity with locked down permissions, which can be reused across task definitions, making role-based access control (RBAC) policy enforcement easier. | Native storage feature |
| Service charge | Free (transactions and early delete charges apply) | Paid. Service fees are described here. Pay for the storage, listing, tiering, and other transactions, and any early deletion penalties. |
Monitoring fee described here. No extra charges for tier transitions, early deletion, or data retrieval |
| Predicates | Time-based, path prefixes, and tags | Object and container metadata with support for parameterization plus path prefixes exclusion | N/A. Smart tiering applies to all objects in the account that infers their tier. |
| Monitoring | Basic - storage logs and Event Grid | Advanced – Dashboard, Detailed reports, Conditions Preview | Storage Logs, Free metrics for each smart tier enabled storage accounts for both object count and size distribution across smart tier |
| Operations | Tiering, deletion | Multi-operation. Full list of supported operations | N/A |
Known limitations
- Tiering isn't supported for append blobs and page blobs.
- Tiering to or from premium accounts isn't supported.
- Smart tiering is account-level only and can only manage objects that infer the tier from the storage account. Objects with explicit tiers set aren't managed. Learn more.
- Smart tiering doesn't delete objects.
- Azure Blob Storage lifecycle management can't act on smart-tier-managed objects.
- Storage Actions and Azure Blob Storage lifecycle management performance might vary.