# Azure Advisor vs ZopNight: Cost Recommendations Compared
> Azure Advisor is free and flags idle VMs, unattached disks and reservation buys. How its checks and lookback windows work, and what ZopNight adds on top.
Source: https://zop.dev/compare/azure-advisor-vs-zopnight
Published: 2026-09-25 · Author: avinash-gaurav · Tags: zopnight, comparison, azure-advisor
---
Every Azure subscription already has Advisor. Before paying for anything else, it is worth knowing exactly how its cost checks decide what to flag.
## What Azure Advisor is
Advisor is Microsoft's built-in recommendation service. Its dashboard sorts findings into five categories: Reliability, Security, Performance, Cost and Operational excellence ([Advisor overview](https://learn.microsoft.com/en-us/azure/advisor/advisor-overview)). The product page says "Azure Advisor is available at no additional cost", with the caveat that acting on some recommendations can add charges ([Azure Advisor](https://azure.microsoft.com/en-us/products/advisor)).
## How the VM cost checks work
Microsoft documents the VM and scale set logic in detail ([cost recommendations](https://learn.microsoft.com/en-us/azure/advisor/advisor-cost-recommendations)):
- **Shut down.** Advisor looks for machines that "weren't used at all over the last seven days". It uses CPU and outbound network only; memory is not considered. A shutdown recommendation needs P95 of maximum CPU below 3%, P100 of average CPU over the last 3 days at or below 2%, and outbound network below 2%.
- **Resize.** This check adds memory. For user-facing workloads the target SKU must keep P95 CPU and network at 40% or lower and P99 memory at 60% or lower. For other workloads both limits rise to 80%. Advisor can suggest a newer generation or a different family, not only a smaller size in the same family.
- **Burstable.** Low-average, spiky workloads can be pointed at B-series SKUs.
- **Lookback.** The default is 7 days. You can choose 7, 14, 21, 30, 60 or 90 days, and a change can take up to 48 hours to show.
Microsoft also lists two limits worth knowing. Savings are "based on retail rates" and ignore your discounts. And the recommendations "don't take into account the presence of Reserved Instances (RI) / Savings plan purchases", so a cross-series resize can even raise costs.
Beyond VMs, the [cost reference](https://learn.microsoft.com/en-us/azure/advisor/advisor-reference-cost-recommendations) covers App Service plans, Cosmos DB, Data Explorer, Databricks autoscaling, AKS, Log Analytics, unattached disks, and reservation and savings plan purchases across many services.
## Where ZopNight differs
Advisor tells you what to change inside Azure. ZopNight works across AWS, GCP and Azure, and it is built to act on what it finds, not only list it.
- **Schedules, not just shutdowns.** Advisor's shutdown check targets machines that did nothing all week. A dev VM busy from 9 to 6 on weekdays never meets that bar. ZopNight's schedule recommendations look for running hours that could be removed, using the latest 21 days of metrics. Acting on one opens a schedule with the resource already attached. See Azure VM scheduling.
- **Deallocate, not just stop.** ZopNight's scheduler handles Azure's deallocate rather than a plain stop. Microsoft bills a Stopped (Allocated) VM but not a Deallocated one, although disks attached to it are still charged ([VM states and billing](https://learn.microsoft.com/en-us/azure/virtual-machines/states-billing)).
- **Your real rate.** Once billing sync is on, ZopNight reads Azure Cost Management using amortized cost, so reservation and savings plan purchases are spread across their term per resource. Schedule savings for billing-connected orgs use the resource's last 30 days of billed cost divided by its measured running hours.
- **Doing the fix.** Supported recommendations have a Remediate button that runs a check, an optional admin approval for destructive steps, the cloud action, and a validation step. Others can be sent to Jira. More in auto-remediation.
- **Checking the fix.** When someone marks a recommendation applied, ZopNight compares the resource against a baseline captured when the recommendation was raised, so "applied" is measured, not just claimed.
| | ZopNight | Azure Advisor |
|---|---|---|
| Clouds | AWS, GCP, Azure | Azure |
| Idle VM rule | Idle, rightsizing and schedule recommendations | Shutdown if unused for the lookback period (default 7 days) |
| Savings basis | Billed cost once billing sync is set up, amortized on Azure | Retail rates; ignores RIs and savings plans |
| Acting on a finding | Remediation wizard, schedule attach, Jira ticket | Recommended action links in the portal |
| Price | See zop.dev/pricing | No additional cost |
## Choose Azure Advisor if...
- You are Azure-only and want a free first pass over VMs, disks and reservations.
- You want reliability, security and performance recommendations in the same place as cost.
- Your savings are mostly one-off: deleting abandoned resources and buying reservations.
## Choose ZopNight if...
- You run AWS or GCP as well, and want one recommendation list with evidence for each item: the monitoring values that triggered the rule.
- Your waste is part-time: VMs, AKS clusters and flexible servers that are busy by day and idle at night. See AKS scheduling.
- You want savings stated against what you actually pay, including reservations, rather than retail rates.
## Frequently asked questions
### Is Azure Advisor free?
Yes. Microsoft's product page says Azure Advisor is available at no additional cost, though acting on some recommendations, such as enabling VM backup, can add charges.
### How far back does Azure Advisor look for idle VMs?
Seven days by default. Microsoft's docs say you can change the lookback to 7, 14, 21, 30, 60 or 90 days, and that updated recommendations can take up to 48 hours to appear.
### Should I turn Advisor off if I use ZopNight?
No. Advisor costs nothing and covers reliability, security and performance as well as cost. ZopNight adds multi-cloud coverage, scheduling, billed-cost savings and remediation. Treat any overlap in cost findings as a second opinion.