Azure Blob Storage Life-cycle Management

In this post, we will talk about a mechanism how we can move data from Azure Storage from one tier to another.

Business needs
Nowadays, Azure is more than a location where you scale your processing power. Slowly, Azure becomes a location where you store your data, including audit data for 2-5 or even 15 years.

Problem
Initially, you have warm data, that needs to be often accessed, for example, audit data for the last 30 days. After a while, this data is becoming cooler, and you don't need to access so often.
Azure Storage has at this moment three different types of tiers (Archive, Cool, Hot) that offers different access speed, but also at the same time, the price is different. For example the storage price between Hot and Archive is 10x different.

Current solution
At this moment even if you have all these three tiers, you had to do the chancing manually. It means that you had to create an automation mechanism that can decide if a specific content/blob can be retired and to do the actual tier change of the storage.
It involves custom development that of course comes with extra costs from running and maintenance perspective.

Lifecycles Management
A new feature of Azure Storage is allowing us to define a custom policy that can move data from one tier to another or execute custom actions. A policy is applied automatically for all blobs that respect a specific rule at the blob level (under a particular container).
For example from a simple policy, we can specify that all blobs under a specific Azure Storage to be kept on the Hot tier for 90 days, moves to Cool storage for the next 180 days and kept for the next 5 years on Archive tier. After this period the content can be automatically removed.

In the below example we define a policy that is applicable for all blobs that have prefix 'autditlogs'  or 'auditdata' with the rules that were specified in the previous paragraph. 

{
  "version": "0.5",
  "rules": [ 
    {
      "name": "auditPolicy", 
      "type": "Lifecycle", 
      "definition": {
        "filters": {
          "blobTypes": [ "blockBlob" ],
          "prefixMatch": [ "auditlogs" , "auditdata" ]
        },
        "actions": {
          "baseBlob": {
            "tierToCool": { "daysAfterModificationGreaterThan": 90 },
            "tierToArchive": { "daysAfterModificationGreaterThan": 180 },
            "delete": { "daysAfterModificationGreaterThan": 2555 }
          },
          "snapshot": {
            "delete": { "daysAfterCreationGreaterThan": 2555 }
          }
        }
      }
    }
  ]
}

Things that you need to consider
When you need to apply a rule immediately, you need to set the number of days to 0 (ZERO).

The below constraints are only for the preview phase. Things might change once it will be in GA.

1. Only 'blockBlobs' are supported for now
2. Snapshots cannot be moved from one tier to another. Only delete operations are supported on base blob and on snapshot also
3. The condition values that are supported are age from last modification or from the creation date

Conclusion

The lifecycle management feature can be seen as a small change, but behind the scene, Azure is closer to enterprise needs and requirements. In this context, being able to define such policies can be a money saver not only from storage cost but also from management cost.