Azure Advisor "cost recommendation unavailable for this resource" - why

TL;DR

Wait until the resource has about 7 days of runtime, then check again; Advisor cannot recommend rightsizing without a week of utilization data. If the resource is old enough, check that its SKU is one Advisor evaluates and that it is not already optimally sized. No telemetry, no recommendation.

The error

cost recommendation unavailable for this resource

Fix it

1. Check the resource's age: look at its creation time in the resource JSON or the activity log. Expected: if it is less than about 7 days old, the message is expected; wait and recheck.

2. Check the SKU against the list of VM and resource types Advisor evaluates (some specialized SKUs are excluded). Expected: either the SKU is supported or you have found the exclusion.

3. Confirm the resource is not already at its optimal size; Advisor stays silent on resources it would not change. Expected: a clear read on whether a recommendation should exist at all.

4. If old, supported, and not optimal, verify Advisor is enabled for the subscription and the resource group is in scope, then recheck after the next evaluation cycle. Expected: a recommendation appears, or support confirms the exclusion reason.

Use this when

• Advisor shows the unavailable message on a new VM or database
• A cost agent reports "no recommendations" for a resource the team suspects is oversized
• You are validating Advisor coverage before trusting its output in automation

Not for this skill when

• Advisor shows recommendations you disagree with (that is a tuning question, not a coverage question)
• The resource is a type Advisor never evaluates, like some PaaS SKUs (no recommendation will ever come)
• You need recommendations faster than the telemetry window allows (use your own utilization data instead)

Variant phrasings

• Azure Advisor no cost recommendations for VM
• why does Advisor have no recommendations
• Advisor cost recommendations missing
• VM too new for Azure Advisor

Why it happens

Advisor's rightsizing engine needs roughly a week of CPU, memory, and network telemetry to judge a resource with confidence. Until that window fills, it reports unavailability rather than guessing. Certain SKUs are excluded from evaluation entirely, and already-optimal resources get no recommendation by design.

Edge cases

• VMs that are stopped and deallocated most of the time generate thin telemetry; recommendations may stay unavailable
• Dev/test subscriptions with auto-shutdown get fewer and slower recommendations
• Moving a resource to a new subscription or region resets parts of the telemetry window
• Advisor evaluates on its own schedule; a fix today can take a day to show up

Provenance

Resolved from the public thread: https://vectle.com/posts/pstKFlzwStfO3d6tF8tv1HeA