Agent Foskett Academy • Microsoft Sentinel • Module 5 • Lesson 47

Lesson 47 — Building Your First Operational Workbook

Last lesson, we designed the workbook.

Now Agent Foskett gets his hands dirty and builds an Identity Sign-In Operations workbook from scratch.

Module 5 — Workbooks and Dashboards • Hands-on build
Agent Foskett building a Microsoft Sentinel operational workbook
What you will build

Identity Sign-In Operations

✓ Failed sign-in trend
✓ Ranked account table
✓ Event-level evidence
✓ Time filters and validation

What you will build

In Lesson 46, we designed an operational workbook. Now we will build one: Identity Sign-In Operations, a reusable view of sign-in activity that helps analysts find unusual patterns and pivot into the underlying records.

The lab uses Microsoft Entra sign-in logs in a Log Analytics workspace. You will create a summary, a trend, a ranked account table and a detail view.

Learning objectives

  • Open an editable Azure Monitor / Microsoft Sentinel workbook.
  • Check the correct workspace and data source before writing queries.
  • Create an hourly sign-in trend using KQL.
  • Add a ranked table of accounts with repeated failures.
  • Add a time-range parameter and apply it consistently.
  • Test the workbook for empty results, performance and permissions.
  • Save and share the workbook with appropriate access.

Prerequisites and safety

  • A Microsoft Sentinel-enabled Log Analytics workspace or an appropriate Azure Monitor workspace.
  • The SigninLogs table populated by a suitable Entra sign-in log diagnostic configuration or connector.
  • Permission to read the workspace data and create/edit the workbook in the intended resource scope.
  • A training or authorised environment. Never use sensitive production identity data in screenshots or demonstrations.

Portal paths and labels can vary as Microsoft Sentinel capabilities transition toward the Defender portal. This walkthrough uses the Azure workbook authoring experience; use the equivalent supported editor in your environment.

Our finished workbook layout

IDENTITY SIGN-IN OPERATIONS [Time Range: Last 24 hours] +------------------------------------+ | Failed sign-ins: hourly trend | +------------------------------------+ | Accounts with repeated failures | +------------------------------------+ | Recent failed sign-in detail | +------------------------------------+ Analyst notes: what this view can and cannot establish

Step 1 — Open the workbook editor

  1. In the Azure portal, open the appropriate Microsoft Sentinel workspace or Azure Monitor workbook area.
  2. Go to Workbooks and choose the option to add or create a workbook.
  3. Open an editable blank workbook, or start from an appropriate blank template.
  4. Confirm you are editing the workbook in the intended subscription, resource group and workspace context.

If your organisation uses the Defender portal for Sentinel operations, follow its current workbook navigation or open the Azure workbook authoring surface provided for your environment.

Step 2 — Give the workbook a purpose

Add a text section with a clear title and one-sentence explanation:

Identity Sign-In Operations Purpose: Help analysts identify unusual sign-in failure patterns and investigate accounts requiring further review. Scope: Entra sign-in events available in this workspace; not all identity activity across the tenant.

This small description prevents future readers from mistaking the workbook for an organisation-wide security verdict.

Step 3 — Check the data source

Before adding charts, confirm that SigninLogs exists and contains recent records. Select the correct Logs / Log Analytics data source and workspace in the workbook query editor.

Data check — are sign-in records available?
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
SigninLogs
| where TimeGenerated > ago(24h)
| summarize Events = count(),
            LatestEvent = max(TimeGenerated)

A zero count, empty latest timestamp or query error requires investigation. Check connector configuration, workspace selection, permissions, ingestion latency and the actual period covered by your data.

Step 4 — Add the first query component

  1. In workbook edit mode, choose Add query (or the equivalent query component).
  2. Choose a Log Analytics / Logs source and the workspace containing SigninLogs.
  3. Paste the hourly trend query below.
  4. Run it and inspect the returned columns before changing the visualisation.

Step 5 — Build the hourly trend

Chart 1 — failed sign-ins by hour
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
SigninLogs
| where TimeGenerated > ago(24h)
| where ResultType != "0"
| summarize FailedSignIns = count()
    by bin(TimeGenerated, 1h)
| order by TimeGenerated asc

Configure a time chart or equivalent time-series display, with TimeGenerated as the horizontal axis and FailedSignIns as the numeric value. Label the component Failed sign-ins by hour.

This counts failed sign-in records as defined by the selected ResultType filter. It does not count unique attackers or compromised accounts.

Step 6 — Read the trend like an analyst

Look for an abrupt increase, repeated activity over several hours or a change from the normal pattern. But ask whether a scheduled job, misconfigured application or authentication outage could explain it.

A spike is a reason to investigate, not proof of malicious activity.

Step 7 — Add the ranked accounts table

Table 2 — accounts with repeated failures
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
SigninLogs
| where TimeGenerated > ago(24h)
| where ResultType != "0"
| summarize Failures = count(),
            UniqueIPs = dcount(IPAddress),
            LastSeen = max(TimeGenerated)
    by UserPrincipalName
| where Failures >= 5
| order by Failures desc

Add a second query component. Display it as a table with account, failure count, distinct source IP count and last-seen time. The threshold of five is for this lab only; it is not a recommended detection threshold.

Step 8 — Add an evidence-level table

Table 3 — individual failed sign-in records
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
SigninLogs
| where TimeGenerated > ago(24h)
| where ResultType != "0"
| project TimeGenerated, UserPrincipalName,
          IPAddress, AppDisplayName, ResultType,
          ConditionalAccessStatus
| order by TimeGenerated desc
| take 100

Add a third query component, displayed as a table. It gives analysts actual events to inspect rather than just summary counts. The take 100 cap limits visible rows; it does not imply there were only 100 failures.

Step 9 — Add a time-range parameter

  1. In edit mode, choose Add parameter or the equivalent parameter control.
  2. Create a time-range parameter, for example TimeRange.
  3. Give it a practical default such as Last 24 hours.
  4. Configure each query component to use the parameter as its workbook time-range setting.
  5. Remove or adapt the hard-coded ago(24h) filters in your KQL so the selected range actually controls the results.

Depending on the workbook editor and query source, the time-range control may be applied through the component’s query settings or by referencing the parameter in KQL. Do not apply a parameter and leave contradictory hard-coded time filters.

Step 10 — Verify the time filter works

Switch between a short and long period and check that the counts and plotted intervals change as expected. For queries with explicit KQL time filters, validate that they use the selected workbook parameter.

For a real deployment, document the chosen time-range binding and test it with your portal version; do not assume a visible dropdown is wired to every query.

Step 11 — Keep the workbook readable

ComponentDisplayAnalyst question
Title and scopeTextWhat does this workbook cover?
Hourly failuresTime-series chartWhen did activity change?
Repeated failuresRanked tableWhich accounts deserve review?
Recent eventsDetail tableWhat actually happened?
Data notesTextWhat might be missing?

Step 12 — Consider an account filter

An account parameter can help analysts narrow a detailed query to a single user. Add it only after the core workbook works, and make sure the query handles a blank or all-accounts selection correctly.

For training, it is perfectly acceptable to start with a time parameter and defer more advanced filters to a later lesson.

Step 13 — Test the empty state

Select a time range with no events, if practical. Does the workbook clearly show an empty result rather than suggesting everything is secure?

Check whether the underlying table is absent, the connector stopped sending data, the selected workspace is wrong, or the query genuinely has no matching rows.

Step 14 — Check data freshness

Freshness check — last available event
  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
SigninLogs
| summarize LatestEvent = max(TimeGenerated),
            TotalRecords = count()
| extend MinutesSinceLatest =
    datetime_diff("minute", now(), LatestEvent)

A stale latest timestamp can make a dashboard misleading. Interpret freshness against expected connector latency and collection configuration. A missing latest timestamp is a data-availability issue, not a clean bill of health.

Step 15 — Test performance

  • Use bounded time ranges and sensible result limits.
  • Summarise high-volume data before plotting it.
  • Avoid duplicating expensive queries across components.
  • Test a longer time range to find slow components.
  • Avoid putting large unfiltered event tables at the top of the page.

Step 16 — Save the workbook

  1. Use the workbook editor’s Save or Save as action.
  2. Choose the correct subscription, resource group, name and location as prompted.
  3. Use a descriptive name such as Identity Sign-In Operations.
  4. Save and reopen the workbook to confirm that the layout, queries and parameters persist.

Saving creates or updates a workbook resource. It does not automatically grant every SOC analyst permission to view its data.

Step 17 — Validate access

Check access using an appropriately authorised analyst identity. Workbook access and underlying workspace query permissions are separate considerations. An analyst may be able to open the workbook but lack permission to run one or more queries.

Follow organisational RBAC and resource-sharing standards. Avoid making sensitive workbooks broadly accessible simply to fix a permission error.

Step 18 — Document assumptions

  • Source: SigninLogs in the selected workspace.
  • Failures: events matching the chosen ResultType condition.
  • Account ranking: at least five matching records in the selected period.
  • Time: workbook time-range selection once wired and tested.
  • Limitations: ingestion delays, missing sources, service accounts and legitimate failures.

Common build problems

SymptomFirst checks
Query returns no rowsTable, connector, workspace, period, permissions
Chart does not renderTimestamp column, numeric value, visual type
Time filter has no effectParameter binding and hard-coded filters
Workbook is slowTime window, result size, repeated queries
Colleague sees an errorWorkbook RBAC and data-source permissions
Counts look inconsistentFilter logic, data latency, query time range

Hands-on challenge

  1. Build the workbook using the three provided queries.
  2. Give every chart or table a meaningful title.
  3. Add and test a time-range parameter.
  4. Record the newest available sign-in event.
  5. Explain why the top-ranked account is not necessarily compromised.
  6. Save, reopen and test access with an authorised analyst.
  7. Write two limitations readers should see before relying on the results.

Agent Foskett takeaway

Start with a question. Build the query. Check the evidence. Then choose the visual.

A dashboard should help the analyst find the next lead — not pretend to have finished the investigation.

The logs already knew.

Lesson 47 complete
You have built an operational workbook with identity queries, meaningful visualisations, evidence-level tables, a time-range control and a repeatable validation checklist.
Sentinel Academy Home

Related Agent Foskett learning

Build on the design principles and earlier workbook introduction.