Hot, cool, cold, archive: Azure Blob storage tiers without overpaying
Storing a file you touch daily and a file you touch once a year at the same price is money on fire. Blob access tiers let you match cost to how often data is actually read — as long as you know the one tradeoff and the one trap.
Not all data is equal. This month's active dataset earns its keep; the compliance archive from four years ago just sits there. Azure Blob Storage lets you price them differently through access tiers, and the whole system runs on one seesaw: the cooler the tier, the cheaper it is to store and the more expensive it is to read. Get the match right and the same data costs a fraction; get it wrong and you either overpay for storage or get stung on retrieval.
The four tiers
| Tier | For data that is… | Storage cost | Access cost | Min. retention | Retrieval |
|---|---|---|---|---|---|
| Hot | accessed frequently | Highest | Lowest | — | Instant |
| Cool | infrequently accessed | Lower | Higher | 30 days | Instant |
| Cold | rarely accessed, still fast | Lower still | Higher still | 90 days | Instant |
| Archive | almost never accessed | Lowest | Highest | 180 days | Hours (rehydrate) |
Hot, cool, and cold are all online tiers — data is readable immediately, in milliseconds, at the same durability. They differ only on the cost seesaw and on a minimum retention period. Hot is for data in active use; cool for short-term backups and older sets you still occasionally read; cold for data you rarely touch but must be able to pull back instantly.
Archive is different — and this is the trap
Archive is not just "very cold." It is an offline tier, and that changes the rules. Storage is the cheapest Azure offers, but you cannot read an archived blob at all without first rehydrating it to an online tier — an operation that can take up to 15 hours. So archive is perfect for long-term backups, raw data you must preserve, and compliance sets that essentially never get read — and completely wrong for anything a user might request and expect to see soon. "Cheapest storage" hides "slowest and most expensive to get back."
The cooler the tier, the less you pay to keep it and the more you pay — in money and in time — to read it.
Each cooler tier carries a minimum retention period — 30 days for cool, 90 for cold, 180 for archive. Delete, overwrite, or move a blob to another tier before that window is up and you pay a prorated early-deletion fee for the days remaining. Drop a file into archive and pull it out after 45 days, and you are billed as if it stayed 180. The tiers reward putting data where it genuinely belongs, not shuffling it around impatiently.
Do not tier by hand — automate it
The real win is not choosing a tier once; it is letting data cool as it ages. That is what lifecycle management policies do: rule-based transitions like "move blobs to cool after 30 days without access, to archive after 180, and delete after 7 years." You set the rules once and Azure moves data down the seesaw automatically, so nobody has to remember. (One caveat worth knowing: a lifecycle policy can move blobs into archive, but it cannot rehydrate them back out — pulling data out of archive is always a deliberate act.)
How to choose
- In active use? Hot. Do not get clever — the access costs of a cooler tier will erase the savings if you are reading it often.
- Occasionally read, kept for a while? Cool or cold, depending on how rare "occasionally" is.
- Backup, compliance, raw archive you essentially never read? Archive — and accept that getting it back takes hours and planning.
- Not sure, and it ages predictably? Put it in hot and write a lifecycle policy to cool it over time.
Match the tier to the read pattern, respect the retention windows, and let lifecycle rules do the moving. That is how a storage bill stops being a flat fee for everything and starts reflecting what your data is actually worth to you this month.