feat(azure compute): populate vCPU + MemoryGB via cached SKU catalogue (closes #148)

[cristim]

Summary

Wires Azure compute to the lazy SKU-catalogue lookup pattern from PR #81, so convertAzureVMRecommendation now populates common.ComputeDetails.VCPU and common.ComputeDetails.MemoryGB from armcompute.ResourceSKUsClient.NewListPager capabilities. Both prerequisites are now in place: PR #97 added the destination fields (with omitempty JSON tags), and PR #81 landed the cache pattern for the other three Azure services.

Closes #148.

Why

common.ComputeDetails.GetDetailDescription() already reads VCPU and MemoryGB and appends " (<vcpu> vCPU / <memory> GB)" when both are >0. Pre-PR, Azure VM recommendations always had both at 0, so the size string was suppressed and the UI rendered SKU names alone (e.g. Standard_D2s_v3). Post-PR they render as Standard_D2s_v3 (2 vCPU / 8 GB).

What changed

• providers/azure/services/compute/client.go:

  • New unexported vmSKUEntry{vCPUs, memoryGB}.
  • New skuCacheOnce sync.Once + skuCacheMap map[string]vmSKUEntry on ComputeClient.
  • cachedSKULookup(ctx, skuName) — lazy single-fetch, O(1) lookup thereafter.
  • fetchSKUCatalogue(ctx) — reuses the existing createResourceSKUsPager and isAvailableInRegion helpers (already used by GetValidResourceTypes); returns nil on error.
  • populateVMSKUMapFromPage extracted out of fetchSKUCatalogue to stay under the project's gocyclo=10 threshold (matches the cache/database extraction pattern from PR perf(azure): batched SKU catalogue lookup eliminates N+1 in recommendation converters #81).
  • extractVMSKUCapabilities parses vCPUs (Atoi) and MemoryGB (ParseFloat) from the SKU's name/value capability list.
  • convertAzureVMRecommendation takes ctx (was _) and populates Details.VCPU/MemoryGB from the cache when both >0. Single caller GetRecommendations already passes ctx.

• providers/azure/services/compute/client_test.go: 4 new tests

  • _PopulatesVCPUAndMemoryFromSKUCache — catalogue hit → fields enriched.
  • _PagerErrorFallsBack — fetch error → fields stay 0; conversion succeeds.
  • _NoMatchLeavesFieldsZero — catalogue miss → fields stay 0.
  • _CachedSKULookup_FetchedOnce — 10 lookups → 1 NextPage call (pins the N+1 invariant).

The shared mocks.MockResourceSKUsPager doesn't expose its page counter and can't simulate a NextPage error, so the tests use a small file-scoped vmSKUCatalogueMockPager (matches the cosmosdb/cache test convention where each service defines its own catalogue mock). No change to the shared mocks surface.

Graceful-degradation contract

A transient ResourceSKUsClient failure no longer breaks Azure VM recommendation conversion: VCPU/MemoryGB stay at 0 (the omitempty JSON tags hide them from API payloads), the rest of Details is populated from the recommendation payload as before, and a WARN is logged once per client lifetime. Matches the contract from cache/cosmosdb/database in PR #81.

Out of scope

• AWS / GCP compute converter wiring (separate follow-ups, per the issue body).
• Platform / Tenancy / Scope enrichment — those need additional Azure data sources beyond armcompute.ResourceSKU.Capabilities.
• No schema change to common.ComputeDetails (fields already added in PR perf(common): extend ComputeDetails with VCPU + MemoryGB #97).

Test plan

• go build ./...
• go test github.com/LeanerCloud/CUDly/providers/azure/services/compute — 42 passed (4 new + all existing)
• go test github.com/LeanerCloud/CUDly/providers/azure/... — 309 passed across 10 packages (the pre-existing providers/azure/services/search build failure is unrelated — go.sum drift)
• go test ./... from pkg/ — 319 passed
• go vet ./... — clean
• gofmt -l . — empty
• gocyclo -over 10 on the modified file — no offenders

Summary by CodeRabbit

• Enhancements
  • Azure VM recommendations now include detailed vCPU and memory capacity information.
  • Improved performance through intelligent caching of Azure VM specification data.
  • Enhanced reliability with graceful fallback handling when specification data is unavailable.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: bc167cdd-870f-4e0e-a5d3-34ef9e8696b4

📥 Commits

Reviewing files that changed from the base of the PR and between c84fd02 and aca5938.

📒 Files selected for processing (2)

• providers/azure/services/compute/client.go
• providers/azure/services/compute/client_test.go

---

📝 Walkthrough

Walkthrough

Azure Compute recommendations are enriched with vCPU and MemoryGB data via a lazily-cached SKU catalogue. The ComputeClient fetches ARM Resource SKUs once per lifetime, parses capabilities, and the recommendation converter populates ComputeDetails from the cache with graceful fallback to zero on failures.

Changes

Azure SKU-Based Compute Enrichment

Layer / File(s)	Summary
Data Model providers/azure/services/compute/client.go	vmSKUEntry struct holds parsed vCPUs and memoryGB; ComputeClient gains skuCacheOnce and skuCacheMap fields for per-client-lifetime cache.
Cache Initialization & Lookup providers/azure/services/compute/client.go	fetchSKUCatalogue fetches the ARM SKU catalogue once via pager, populateVMSKUMapFromPage extracts region-specific VM SKUs, and extractVMSKUCapabilities parses vCPU/MemoryGB from capability pairs; cachedSKULookup wraps with sync.Once and returns (entry, hit) or (empty, false).
Recommendation Conversion providers/azure/services/compute/client.go	convertAzureVMRecommendation enriches ComputeDetails with InstanceType, VCPU, and MemoryGB from the SKU cache; logs warnings and degrades to zero on cache miss or fetch error.
Tests & Helpers providers/azure/services/compute/client_test.go	Mock pager vmSKUCatalogueMockPager and buildVMSKU helper support four tests: SKU cache population, pager error fallback, missing SKU fallback, and single-fetch-per-lifetime invariant.

Sequence Diagram

sequenceDiagram
    participant Converter as convertAzureVMRecommendation
    participant Client as ComputeClient
    participant Cache as SKU Cache<br/>(sync.Once)
    participant Pager as ResourceSKUsClient<br/>Pager
    participant ARM as Azure SKUs<br/>API

    Note over Converter,ARM: First recommendation conversion
    Converter->>Client: cachedSKULookup(vmName)
    Client->>Cache: Check cache (sync.Once gate)
    alt Cache not yet initialized
        Cache->>Pager: fetchSKUCatalogue()
        Pager->>ARM: ListByLocation(region) paginated
        ARM-->>Pager: SKU pages
        Pager->>Cache: Build name→{vCPU, MemoryGB} map
        Cache-->>Client: Return (entry, hit)
    else Cache already initialized
        Cache-->>Client: Return (entry, hit)
    end
    Client-->>Converter: {vCPUs, MemoryGB}
    Converter->>Converter: Populate ComputeDetails

    Note over Converter,ARM: Second recommendation (same client)
    Converter->>Client: cachedSKULookup(otherVmName)
    Client->>Cache: Check cache (sync.Once skipped)
    Cache-->>Client: Return from map (no API call)
    Client-->>Converter: {vCPUs, MemoryGB}
    Converter->>Converter: Populate ComputeDetails

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related issues

• perf(azure): batched SKU catalogue lookup to populate deferred Details fields without N+1 API calls #49 — Parent issue scoping SKU catalogue enrichment across cache/cosmos/database/compute; this PR closes the compute variant deferred in that scope.
• perf(azure): batched SKU catalogue lookup eliminates N+1 in recommendation converters #81 — Establishes the lazy sync.Once caching pattern (cachedXxxLookup, fetchXxxCatalogue) that this PR mirrors for Azure compute SKU catalogue.
• perf(aws): populate ComputeDetails VCPU/MemoryGB from EC2 DescribeInstanceTypes catalogue #99 — AWS counterpart implementing vCPU/MemoryGB enrichment via DescribeInstanceTypes, demonstrating the same cross-provider enrichment strategy.

Possibly related PRs

• perf(common): extend ComputeDetails with VCPU + MemoryGB #97 — Extends common.ComputeDetails struct with VCPU and MemoryGB fields that this PR populates on the Azure side via SKU cache lookup.

Suggested labels

triaged, priority/p3, severity/low, impact/few, type/feat

Poem

🐰 A rabbit caches Azure SKUs,
One fetch per lifetime—no N+1 blues!
vCPUs and memory now brightly glow,
Recommendations enriched, fast and low.
Hops through the pager, once and for all! 💾

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The pull request title clearly describes the primary change: populating vCPU and MemoryGB fields in Azure compute recommendations using a cached SKU catalogue, and correctly references the closing issue #148.
Linked Issues check	✅ Passed	The PR successfully implements the Step 3 (Wire Azure) portion of issue #148: it adds cachedSKULookup pattern, wires convertAzureVMRecommendation to populate VCPU/MemoryGB from the SKU catalogue with graceful degradation, and includes comprehensive tests for SKU cache hits, errors, misses, and fetch-once invariant.
Out of Scope Changes check	✅ Passed	All changes align with the Azure wiring scope of issue #148; the PR explicitly defers AWS/GCP converter updates and other cross-cutting work to follow-ups, maintaining focused scope.
Docstring Coverage	✅ Passed	Docstring coverage is 85.71% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/issue-148-azure-compute-vcpu-memory

---

_{Review rate limit: 2/5 reviews remaining, refill in 29 minutes and 19 seconds.}

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[cristim]

@coderabbitai review

[coderabbitai]

✅ Actions performed

Review triggered.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

@coderabbitai review

[coderabbitai]

✅ Actions performed

Review triggered.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

@coderabbitai review

[coderabbitai]

✅ Actions performed

Review triggered.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.