Hot vs Cool vs Archive Storage Tiers in Azure Blob Storage

The four tiers at a glance

Hot tier: frequently accessed data

Hot tier is designed for data accessed multiple times per month. It has the highest storage cost ($0.018/GB) but no retrieval charge. Every read operation carries no data retrieval fee — you pay only per-API-operation charges, not per byte of data downloaded.

Use Hot tier for:

• Active application assets — images, CSS, and JavaScript files served to users in real time
• Recent operational logs actively queried by monitoring tools
• Database backup files for recent backups where restore is likely within days
• Data feeding real-time processing pipelines or dashboards
• Files with unpredictable access patterns where retrieval cost uncertainty outweighs savings from a cooler tier

Cool tier: infrequent access

Cool tier is designed for data accessed less than once a month. Storage cost drops to $0.010/GB/month (44% less than Hot), but reads incur a $0.01/GB retrieval fee and a 30-day minimum storage period applies.

The break-even point where Cool becomes cheaper than Hot depends on read frequency. For a 100 GB dataset:

• Hot at 0 reads: $1.80/month
• Cool at 0 reads: $1.00/month — Cool is cheaper
• Cool with 1 full dataset read per month: $1.00 + (100 GB × $0.01) = $2.00 — more expensive than Hot
• Cool with 0.5 reads per month: $1.00 + $0.50 = $1.50 — cheaper than Hot’s $1.80

Cool tier wins when data is accessed less than approximately once per month on average. Suitable for: monthly financial reports, recent backups that are rarely restored, application logs older than 7 days reviewed only during incidents.

Cold tier: rarely accessed, immediate retrieval

Cold tier (added to Azure Blob Storage in 2023) sits between Cool and Archive. Storage cost is $0.0045/GB/month — 75% cheaper than Hot — but retrieval costs $0.03/GB and the minimum storage period is 90 days. Access is still immediate (no rehydration needed), making it practical for data that might occasionally be needed on short notice but is rarely accessed in normal operations.

Cold tier is appropriate for:

• Database backups older than 30 days kept for compliance but rarely restored
• Video content that has moved out of the active catalogue but may still be requested occasionally
• Application logs older than 30 days where occasional investigation queries are expected but uncommon
• Archived project files that need to remain accessible without advance notice

Cold tier is preferred over Archive whenever data might be needed within a few hours. The rehydration delay of Archive makes it unsuitable for any data with an informal SLA of “available today if needed.”

Archive tier: long-term cold storage

Archive tier is the cheapest storage available at $0.00099/GB/month — approximately 18x cheaper than Hot. The trade-off is substantial: archived blobs cannot be read directly. To access an archived blob, you must first rehydrate it — either by copying it to Hot or Cool tier (Copy Blob), or by changing the blob’s tier (Set Blob Tier) — a process that takes 1–15 hours for standard rehydration.

High-priority rehydration is faster (typically under 1 hour for blobs under 10 GB) but costs more: $0.02/GB vs $0.01/GB for standard priority. Until rehydration completes, reading the blob returns HTTP 409 (conflict).

Archive tier is appropriate for:

• Long-term compliance backup retention — backups kept for 7 years that will almost never be restored
• Raw source data before transformation — original files kept after an analytics pipeline has processed them
• Audit logs and compliance records accessed only during legal discovery or periodic audit
• Older versions of large media files superseded by updated formats

Real cost example: 10 TB backup dataset

A company stores 10 TB of backup data. Backups from the past 7 days are routinely restored; older backups are kept but rarely accessed; backups older than 1 year are kept only for 7-year regulatory retention.

If all 10 TB were stored in Hot tier: 10,240 GB × $0.018 = $184.32/month. Mixed tiering reduces this by approximately 78%. For organisations with large backup or archival footprints, lifecycle-based tiering is one of the highest-return cost optimisation actions available.

Automating tier changes with lifecycle management

Azure Blob Storage lifecycle management policies automate tier transitions without manual intervention. Policies can scope to all blobs in an account or to blobs matching a container prefix or index tag filter.

{
  "rules": [
    {
      "name": "backup-tiering",
      "enabled": true,
      "type": "Lifecycle",
      "definition": {
        "filters": {
          "blobTypes": ["blockBlob"],
          "prefixMatch": ["backups/"]
        },
        "actions": {
          "baseBlob": {
            "tierToCool": { "daysAfterModificationGreaterThan": 7 },
            "tierToCold": { "daysAfterModificationGreaterThan": 30 },
            "tierToArchive": { "daysAfterModificationGreaterThan": 365 }
          }
        }
      }
    }
  ]
}

Apply lifecycle policies via the Azure portal, CLI, or Bicep/ARM templates. Policies evaluate daily. Enable last access time tracking on the storage account if you want to tier blobs based on when they were last read — useful for media libraries where modification date and access date diverge significantly. Note that enabling last-access time tracking adds a small per-operation charge for tracking reads.