Key Takeaways from thjs blog -

• Azure Storage Discovery and Azure Storage Insights share a name and almost nothing else. Storage Insights is operational monitoring with 93-day retention; Azure Storage Discovery is estate analytics with up to 18 months.
• Azure Storage Insights displays a maximum of 200 storage accounts, which disqualifies it as an enterprise estate view no matter how many subscriptions you select.
• Azure Storage Discovery pricing runs on two meters (per storage account and per million objects). At 2,000 accounts and 30 billion objects it lands near $6,900 a month, and the object meter dominates at any real scale.
• Blob inventory reports are the only native source of per-blob size, access tier, and last-access time — the raw material for a tiering business case that survives a finance review.
• None of the three reports cost, and none acts on what it finds. Joining capacity to dollars is manual work in every case, and no service here sees a managed disk.

Part 1 mapped Azure’s eight native storage visibility surfaces and found that almost none of them tell you what to do, and that no single service reports capacity and cost together. This part takes the three that actually measure a blob estate (Azure Storage Discovery, Azure Monitor Storage Insights, and Azure blob inventory reports) and works out what each is good at, where each stops, and what Azure Storage Discovery costs once your estate is real.

What Does Azure Storage Discovery Do That Storage Insights Doesn’t?

Azure Storage Discovery answers “what does my whole blob estate look like, and how is it changing?” while Azure Monitor Storage Insights answers “is this storage account healthy right now, and why did it throttle at 03:00?” The names suggest overlap that does not exist. Choosing between them is usually a category error; most estates end up running both, plus blob inventory underneath.

Azure Storage Discovery is a fully managed, agentless analytics service that aggregates capacity, activity and configuration insights across a Blob and Azure Data Lake Storage estate from a single workspace, with retention of up to 18 months on the Standard plan.

Azure Monitor Storage Insights is an Azure Monitor workbook over platform metrics that reports availability, latency, capacity, throttling and transaction volume per storage account, in near real time, with 93-day retention.

An Azure blob inventory report is a scheduled CSV or Parquet file that lists every blob, version, snapshot and container in a storage account with its size, access tier and last-access time. It is raw enumeration with no interpretation attached.

Retention is the cleanest way to tell the first two apart: 93 days versus 18 months. Granularity is the cleanest way to place the third: the other two aggregate, blob inventory enumerates.

What Is Azure Storage Discovery Good At for an Enterprise Blob Estate?

Azure Storage Discovery is good at one job the rest of Azure cannot do: showing how a blob estate of hundreds or thousands of accounts has grown across more than one budget cycle. It reached general availability on 16 October 2025. You create a workspace, point it at subscriptions or resource groups, and it aggregates insights without deploying anything and without generating transactions against the accounts it analyses.

Estate scale. The overview documentation describes a single workspace analysing up to a million storage accounts. Nothing else native comes close, and the 200-account ceiling on Storage Insights is precisely the gap this fills.

Long retention. The Standard plan retains 18 months, which is the first time Azure has offered a native view of storage growth spanning a full year plus the comparison period before it.

Immediate usefulness. On workspace creation, Storage Discovery backfills the prior 15 days, so it is not empty on day one the way most telemetry deployments are.

Zero impact. Microsoft documents no transactions and no performance impact on analysed storage accounts. The DIY alternative, listing blobs at scale with your own scripts, is itself billable and slow.

Configuration insight, not just capacity. It flags public access, shared key authentication, encryption settings, redundancy, and whether lifecycle management is configured, which makes it a governance surface as well as a capacity one.

Copilot integration. Natural-language questions about the estate with no KQL required, which lowers the barrier for finance and platform stakeholders who will never learn a query language.

Where Does Azure Storage Discovery Stop?

Azure Storage Discovery stops at three boundaries: it sees only Blob and ADLS, it reports no cost, and it takes no action. Each one catches teams out in a different way.

Narrow scope. Azure Files, Queues and Tables are all unsupported. This surprises people because Files is a storage account service. Managed disks are outside the storage resource provider entirely and were never in scope.

No cost data. Azure Storage Discovery will tell you an account holds 400 TB in the hot tier with no lifecycle policy. It will not tell you what that costs or what tiering would save. You join to Cost Management yourself, on resource ID.

No remediation. It is read-only. It identifies infrequently accessed data; it does not tier it, delete it or resize anything. The documented path to action is CSV export into your own scripts.

The Free plan is a demo. Fifteen-day retention means no meaningful trend analysis. Anyone evaluating Storage Discovery on the Free plan is evaluating a different product, and Microsoft’s own documentation disagrees with itself about what Free includes (see below).

Aggregation is slow. Initial aggregation “often completes within a few hours but can also take more than a day,” and the same lag applies to newly created accounts and configuration changes. This is not an alerting surface.

Downgrade is destructive and irreversible. Moving a workspace from a paid plan to Free retains only 15 days, and Microsoft states that all previously aggregated insights are permanently deleted and “can’t be recovered, even if you switch the workspace back to a paid plan.” A cost-cutting exercise can silently destroy a year of trend data. Export first.

Quotas are low by default. Ten workspaces per subscription per region, 100 workspace root entries, 10 scopes per workspace, five ARM tags per scope. All raisable by support request, but check them before designing a topology around them.

What Does Azure Storage Discovery Pricing Cost at Enterprise Scale?

Azure Storage Discovery pricing on the Standard plan bills two meters, per storage account analysed and per million objects analysed. A 2,000-account, 30-billion-object estate lands near $6,900 a month at US list rates. Microsoft publishes the rates but not a worked example, and the two-meter structure makes the bill hard to intuit. The Free plan costs nothing.

There is one ambiguity in Microsoft’s published wording, and it changes the answer by up to 30%. The tiers are written in bracket form (“up to 1000 / greater than 1000 and up to 10000”), which does not state whether crossing a threshold re-rates your entire volume at the cheaper rate (flat-rate) or prices only the incremental band (graduated). Microsoft’s only steer is that “the more storage resources you analyze, the more the average resource cost decreases,” which is true under either model.

House convention points to graduated. Microsoft’s Blob Storage cost guidance is explicit that Hot tier capacity works that way: “the price of the first 50 TB and the price of the second 50 TB are factored into this calculation.” Azure Storage Discovery most likely follows suit, but Microsoft has not said so for this meter. Both readings, at three estate shapes:

The object meter dominates at any real scale; in the largest example it is more than three quarters of the bill, so model object count, not account count. At the top end, visibility into a storage estate becomes a six-figure annual line item in its own right. That is not an argument against Storage Discovery. It is an argument for scoping workspaces deliberately rather than pointing one at the whole tenant because it is easy. The two models converge at small scale and diverge sharply at large scale, so if your estate is big enough for the difference to matter, run your exact numbers through the Azure Pricing Calculator rather than trusting either column here.

Which Insights Does the Azure Storage Discovery Free Plan Include?

Three Microsoft sources say the Azure Storage Discovery Free plan includes configuration insights; one says it does not. If you are evaluating on Free, that contradiction matters.

The Storage Discovery pricing page on Microsoft Learn shows Configuration as “N/A” for the Free plan. The azure.com pricing page states the Free plan “includes Capacity and Resource configuration insights.” The Storage Discovery FAQ agrees with azure.com and is the most specific of the three: it says “Activity, Security, and Consumption reports show insights only for Standard pricing plan and not for Free plan,” naming Capacity and Configuration as the two reports that do appear on Free. The GA announcement uses the same framing.

Plan on the assumption that Free includes configuration insights. If that distinction determines your evaluation, create a Free workspace and look rather than trusting any of the four pages.

What Does Azure Monitor Storage Insights Report?

Azure Monitor Storage Insights reports availability, end-to-end and server latency, capacity used by blob, file, table and queue, failures and throttling, and transaction volumes, for a single storage account or estate-wide through Azure Monitor. It is a workbook over platform metrics, which collect by default, so there is nothing to configure, and the workbook itself is free.

It is already on. No agent, no diagnostic setting, no cost. For incident work this matters more than any feature; it is the view an on-call engineer opens first when an account starts throttling.

Real operational depth. Server latency versus end-to-end latency, split by API, is the fastest way to tell whether a slow application is the storage account’s fault or the network’s.

Broader service coverage than Storage Discovery. Azure Storage Insights covers blob, file, queue and table, the broadest storage-native scope in Azure, where Storage Discovery covers Blob and ADLS only. For an Azure Files share, Storage Insights is the only native option in this part.

Where Does Azure Storage Insights Stop?

Azure Storage Insights stops at 93 days and 200 storage accounts, and those two numbers rule it out as an enterprise estate view. It is an incident tool, not a trend tool.

93 days - Azure Monitor platform metrics are retained for 93 days, and a single chart can only query 30 days at a time. Year-over-year growth analysis is impossible without routing metrics to a Log Analytics workspace through diagnostic settings, which costs ingestion and retention money.

200 storage accounts - The multi-account view displays a maximum of 200 accounts regardless of subscription selection. For an enterprise estate this is the disqualifying constraint, and it is precisely the gap Storage Discovery was built to fill.

Ten subscriptions - when customising workbooks, and only three error types shown by default with the remainder bucketed as “Other.”

No cost data - Azure Storage Insights reports capacity and performance, never dollars. Cost lives in Microsoft Cost Management.

What Do Azure Blob Inventory Reports Give You That Aggregates Don’t?

Azure blob inventory reports give you every object in a storage account with its size, access tier and last-access time, per row. That is the one thing neither Azure Storage Discovery (which aggregates) nor Cost Management (which stops at the account) will produce. That makes them the least-discussed service in this part and arguably the most useful for cost work.

On a daily or weekly schedule, Azure writes a CSV or Parquet file into a destination container listing every blob, version, snapshot and container, with a manifest and a checksum file marking completion. Alongside Name, Creation-Time and BlobType, each row carries Content-Length, AccessTier and LastAccessTime.

Size plus tier plus last access, per object, is exactly the input a tiering business case needs. Aggregate services tell you an account “holds infrequently accessed data”; an Azure blob inventory report tells you which 40 TB, under which prefix, untouched since when. Per-object evidence survives the scrutiny that aggregates do not.

One dependency is easy to miss: LastAccessTime is only populated if last-access-time tracking is enabled on the storage account. Enable it before you enable inventory, or the column that carries the business case arrives empty.

What Should You Know Before Enabling Azure Blob Inventory?

Enable Azure blob inventory now, even if nobody analyses the output yet, because reports are not retrospective and the history starts the day you switch them on. Five constraints shape how to deploy them.

Cheap, but not free. Microsoft’s documentation cites $0.003 per million objects scanned, plus normal storage and transaction charges on the generated files. The worked example in the docs comes to a few cents a month.

Slow at scale. Maximum completion time is six days, runs do not overlap, and hierarchical-namespace accounts with hundreds of millions of blobs may take more than a day.

Do not reconcile against the bill. Microsoft warns explicitly that report object counts should not be compared to billed amounts, because inventory excludes metadata and system logs.

Policy limits. 100 rules per policy; cannot write to containers with object replication policies; cannot be configured where version-level immutability is enabled.

Deduplicate. Rare duplicate rows occur; deduplicate on Name plus VersionId before summing Content-Length.

How Do Azure Storage Discovery, Storage Insights and Blob Inventory Compare?

The three services differ on retention, granularity and scale ceiling more than on anything else, and none of them reports cost. The table is the short version of this part.

Which Discovery Layer Tool Should You Use for Each Job?

Use Azure Storage Discovery for estate trend, Azure Storage Insights for incidents, and blob inventory for evidence. Expect to run all three, because they are not substitutes.

‍

What Does Running the Discovery Layer Cost an Enterprise?

Running the discovery layer well costs an enterprise the Storage Discovery Standard subscription, a small inventory charge, optional Log Analytics ingestion, and a recurring engineering cycle to join capacity to cost by hand. The table below is the module to put in front of a CFO when the question is “what does visibility actually cost.”

At the 2,000-account shape, Azure Storage Discovery pricing is roughly $83,000 a year before anyone acts on a single finding, and the engineering cycle that turns its output into a cost decision runs a further one to two engineer-days a month. Blob inventory is effectively free and produces the only per-object evidence in the stack, which makes it the highest-return line in the table by a wide margin.

For a board deck, the whole section reduces to this: the discovery layer tells you what you hold at a cost that scales with object count, and turning that into a dollar decision is a senior engineer’s recurring job, because nothing in the layer prices what it finds.

What Do None of These Three Services Measure?

None of the three services in this part reports what any of the data costs, and none of them acts on what it finds. The specific measurement that is missing sits between them: which containers and prefixes are being read, how that has changed, and therefore what should move tier, observed rather than guessed at in a policy written in advance.

That is the gap Lucidity’s object storage work fills: mapping an estate down to container and prefix level, classifying data by observed access, and flagging accounts whose tier or redundancy no longer matches how the data is used. If you want that view for your own estate before deciding whether it matters, the free storage audit produces it in about 15 minutes without an agent.

                                 Frequently Asked Questions

Should I use Azure Storage Discovery or Azure Storage Insights?

Most estates need both, because they answer different questions. Use Azure Storage Insights for operational work (latency, throttling, errors on a specific account) where near-real-time metrics and zero cost matter. Use Azure Storage Discovery for estate-wide capacity and configuration trend, where 18-month retention and the absence of a 200-account ceiling matter. If you can only run one and the problem is capacity growth across a large estate, Storage Discovery is the correct choice.

How much does Azure Storage Discovery pricing come to for a large estate?

The Standard plan bills on two graduated meters: per storage account analysed and per million objects analysed. A 2,000-account estate holding 30 billion objects comes to roughly $6,900 per month at US list rates. The object meter dominates at scale; in a 150-billion-object estate it accounts for more than three quarters of the bill. Model object count rather than account count when forecasting.

Does Azure Storage Discovery cover Azure Files?

No. Azure Storage Discovery covers Azure Blob Storage and Azure Data Lake Storage only. The deployment planning documentation explicitly lists Azure Files as unsupported, along with Queues and Tables. For Azure Files the native option is Azure Monitor Storage Insights, which covers blob, file, queue and table, with 93-day retention and a 200-account display limit. Managed disks are outside the storage resource provider altogether and are not covered by any service here.

What happens if I downgrade an Azure Storage Discovery workspace to the Free plan?

Retention drops to 15 days and all previously aggregated insights are permanently deleted. Microsoft states explicitly that this data cannot be recovered even if the workspace is switched back to a paid plan. Treat a downgrade as a destructive operation and export anything you need first; CSV export is supported.

What is the best way to identify cold blob data for tiering?

Azure blob inventory reports. Each row carries Content-Length, AccessTier and LastAccessTime per object, which is the only native combination that supports a defensible tiering case. Azure Storage Discovery will flag infrequently accessed data at an aggregate level, which is useful for spotting where to look, but aggregates rarely survive a finance review. Inventory reports are not retrospective and LastAccessTime needs last-access-time tracking enabled, so switch both on before you need the history.

‍