LeftSize vs Azure Advisor
Azure Advisor is Microsoft's built-in recommendation engine for Azure. LeftSize doesn't replace it: it takes the same kind of findings, plus service retirements and governance checks, and puts them in the GitHub repository of the team that owns the resources, as issues with fix steps.
Last checked against Azure Advisor's documentation in October 2026.
| LeftSize | Azure Advisor | |
|---|---|---|
| Where findings appear | GitHub issues in the repository you choose, grouped per scope and updated on every scan | Azure portal (Advisor dashboard), with filters per subscription and resource type |
| Categories | Cost, deprecations and retirements, security and governance: 91 Azure rules | Cost, security, reliability, performance and operational excellence |
| Service retirements | 38 Azure retirement and end-of-support rules, each with the date and a migration path, raised as issues | Service Retirement workbook (preview) in the Advisor workbook gallery |
| Remediation | Issue text with steps, plus @leftsize script for Azure CLI, PowerShell or Terraform, howto, impact and alternatives | Recommended action inline, sometimes a one-click fix or a link to documentation |
| Tracking | GitHub issue history, assignees and labels; snooze or ignore with @leftsize commands | Postpone, dismiss or mark recommendations completed in the portal |
| Clouds | Azure and AWS | Azure only |
| Setup | GitHub App plus a workflow with a read-only (Reader) identity | None; available in the portal for Owner, Contributor and Reader roles |
Where Azure Advisor is the better choice
- Reliability and performance recommendations, which LeftSize does not cover.
- Rightsizing and reservation advice based on usage telemetry across many services.
- Security recommendations from Microsoft Defender for Cloud appear in the same view.
- Nothing to install: it is already in every Azure subscription.
What LeftSize adds
- Findings reach the team that owns the resources, in the repository they already work in, instead of a portal someone has to check.
- Retirements as tracked work: 38 rules for retiring VM sizes, end-of-support runtimes, engines and SKUs, each with its date.
- Remediation scripts and explanations on demand from the issue itself.
- The same workflow for AWS (97 rules), so teams with both clouds have one process.
Using them together
Keep Advisor for reservations, rightsizing and reliability. Use LeftSize to turn waste and retirement findings into GitHub issues for the owning teams, so they are fixed instead of read.