All cheat sheets
Storagecheat sheet

S3 Glacier Instant Retrieval

Archive pricing with millisecond reads: ~68% cheaper storage than S3 Standard, same GET API, no restore jobs — the sweet spot for compliance and archive data you touch a few times a year but need instantly when you do.

Last reviewed: July 11, 2026

TL;DR: Glacier Instant Retrieval is the "have both" tier: storage ~68% cheaper than S3 Standard, but reads are still ordinary millisecond GET calls — no restore jobs, no application changes. The catch is per-GB retrieval fees and a 90-day/128 KB minimum, so it fits data touched a few times a year that must be available immediately when asked for: compliance pulls, legal requests, medical imaging, licensing archives.

The numbers

  • Storage: ~$0.004/GB-month (vs $0.023 Standard, $0.0125 Standard-IA)
  • Reads: normal S3 GETs, milliseconds — but charged ~$0.03/GB retrieved
  • Minimums: 90-day storage duration; objects billed at ≥128 KB
  • Durability: 11 nines across 3+ AZs — identical to Standard

A 10 KB object deleted after 45 days is billed as 128 KB for 90 days — both minimums stack. This class pays off at scale and with patience.

Do this

  1. Find candidates: buckets whose data goes quiet after an active window but where "we'll get it back to you in 4 hours" is unacceptable (regulator requests, customer escalations, clinician access).

  2. Use a lifecycle transition, not manual re-uploads:

    aws s3api put-bucket-lifecycle-configuration --bucket YOUR-BUCKET \
      --lifecycle-configuration '{"Rules":[{"ID":"to-glacier-ir","Status":"Enabled",
      "Filter":{},"Transitions":[{"Days":30,"StorageClass":"GLACIER_IR"}]}]}'
    

    Classic ladder: Standard for days 0–30 → Glacier Instant Retrieval for the 30–365 window → Deep Archive after a year.

  3. Watch retrieval spend for the first few months (Cost Explorer, filtered to the bucket). Climbing retrieval fees mean the data isn't as cold as you thought — move it to Intelligent-Tiering or back to Standard.

  4. Pair with S3 Object Lock where immutability matters — WORM compliance plus instant access plus archive pricing is a hard combination to beat for regulated industries.

Gotchas

  • Retrieval fees are the whole game. Storage math looks amazing until a surprise audit pulls 500 GB. Budget retrieval as a real line item; if access becomes monthly, you're in the wrong class.
  • Tiny objects get rounded up to 128 KB. Millions of small log files will quietly bill several times their actual size — aggregate into archives first.
  • 90-day minimum punishes churn; don't transition anything that might be deleted or rewritten soon.
  • Don't confuse the Glacier siblings: Instant = milliseconds at ~$0.004/GB; Flexible = hours at ~$0.0036/GB; Deep Archive = 12 h+ at ~$0.001/GB. The access pattern, not the price tag, picks the tier.

Skip this if

  • Access is monthly-or-more — Standard or Intelligent-Tiering wins once retrieval fees are counted.
  • You can tolerate hours of wait — Flexible or Deep Archive is meaningfully cheaper.
  • Objects are small (under ~128 KB average) or short-lived (under 90 days) — both minimums will eat the discount.

Run this audit with your AI assistant

Paste this into Claude, ChatGPT, or any agent that can run the AWS CLI with read-only credentials. It audits your account for exactly the waste this sheet describes — and changes nothing.

You are auditing an AWS account for S3 Glacier Instant Retrieval savings.
Use the AWS CLI with READ-ONLY credentials. Do not create, modify, or
delete anything — report findings and recommended (unapplied) fixes only.

1. Per bucket (top 15 by size): storage per class (CloudWatch
   BucketSizeBytes by StorageType), object count, average object size.
2. Candidates for GLACIER_IR: data in STANDARD or STANDARD_IA that is
   rarely read but must stay instantly readable. Use request metrics /
   Storage Lens if enabled; otherwise flag archive/compliance-named
   buckets and prefixes for the user to confirm access patterns.
3. Constraints to check per candidate: 90-day minimum storage duration,
   128 KB minimum billable object size (flag buckets with small average
   objects), and per-GB retrieval fees on GETs (~$0.03/GB) — estimate
   retrieval cost from recent BytesDownloaded if available.
4. Cost math (us-east-1): Standard $0.023/GB-mo, Standard-IA $0.0125,
   Glacier Instant Retrieval $0.004. Compute break-even reads/month per
   candidate: at what monthly GB-retrieved does GLACIER_IR stop winning?

Report a table: bucket/prefix | size | avg object | current class |
est. $/mo now | est. $/mo on GLACIER_IR | monthly-read break-even |
recommendation. Output lifecycle-transition JSON for the clear wins —
but do NOT apply anything.
Works with any assistant that can run shell commands.

Want the guided version?

The S3 Glacier Instant Retrieval walkthrough covers this topic interactively — it asks about your setup, branches to what’s relevant, and quizzes you on the tricky parts. Free and anonymous.

Start the walkthrough