CIS Data Foundations study guide

Start here

CMDB basics

The CMDB is a single system of record for CIs, services, and configured assets, and the hub that Service Management, Asset Management, and Operations Management depend on. A CI can be logical (an application), physical (a server), or an operational construct (a cluster).

The three pillars and the four deployment steps

Two critical success factors: stakeholder alignment on goals, and a strategic connection to both organizational strategy and IT's tactical goals. Plan and Govern are the practices that minimize technical debt, and clear governance plus continuous health monitoring is what makes a CMDB operational.

Teams, governance, and data model design

• Form a Configuration Control Board (CCB): the team that sets direction, decides what the CMDB needs to support strategic initiatives, and decides when it's ready to use. Get executive buy-in early.
• The CMDB team doesn't own the data in the CMDB. Data owners must be part of governance; the CMDB team can usually only run basic checks like duplicates and missing attributes.
• Put in only the data your use cases need. Start with out-of-the-box classes (for example, Firewall, Server, and Application for a security use case), and add classes as new use cases come up.
• For each class you ingest: review its base attributes, identify required and missing attributes, and identify the authoritative source for each attribute. Custom classes and attributes create technical debt, and a CMDB health check looks for them.
• Discoverable attributes (host name, IP address, operating system) come from discovery tools. Non-discoverable ones (groups, owners, location) need manual entry or another source; use Data Manager certification and Health Dashboard recommended fields to find gaps.

Key terms

• Configuration Item [cmdb_ci] is the core CI table that every CI class extends, and it references foundation tables like sys_user and cmn_location. (It extends the Base Configuration Item [cmdb] table.)
• A class is a table of CIs that share attributes. A table can be both a parent and a child (Server extends Computer; Windows Server extends Server). View the structure in CI Class Manager or the schema map.
• Logical CIs take up no physical space and depend on physical components, for example service offerings, services, and business applications. They're entered manually.
• An asset becomes a CI when it's deployed and stays one until it's returned to stock or disposed of. Configuration Management focuses on operational use; IT Asset Management focuses on financial and contractual life cycle.

Asset vs. CI

An asset is often a CI, but a CI isn't necessarily an asset. A business service is a CI with no financial value of its own.

The three "application" records

Foundational data

Foundational data lives in common tables outside the CMDB and is referenced from CI fields: cmn_location, cmn_building, cmn_cost_center, core_company, cmn_department, sys_user, sys_user_group. It usually arrives from LDAP, Active Directory, or HR sources.

Management tools (all free)

CMDB Workspace (from the Store, spans all three pillars), CI Class Manager, CMDB Health Dashboard, CMDB Data Manager, Duplicate CI Remediator, De-duplication Templates, Unified Map, Query Builder, and Intelligent Search are all no-extra-license tools. The CMDB and CSDM Data Foundations Dashboards are a free Store app.

1. Govern the CMDB35% of the exam

CMDB Health Dashboard: three scorecards

• Health rolls up from class level to the overall CMDB, and can also be viewed per CMDB group (use groups of type Health for regional views).
• Health Inclusion Rules are the first filter. They define which classes (and which CIs) count, so a discovery tool that scans everything doesn't drown your score. Use the "is a" operator to include a class hierarchy. A rule can apply to one metric and not another, which is why two metrics can evaluate different CI counts for the same class.
• Scores only appear once the dashboard's scheduled jobs run. Run them as an admin. Job status is in cmdb_health_metric_status.
• Those scheduled jobs aren't active by default; activate them before expecting scores. The dashboard itself is free and opens from CMDB Workspace > Management.
• The dashboard shows a Health Scorecard (overall Completeness, Correctness, Compliance) and a Top 10 view of the classes hurting each score, with drill-down to CIs.
• Create a regional or team view: Configuration > CMDB Groups > New, set Group type = Health, add encoded queries in the CMDB Group Contains Encoded Queries related list, check Show All CI, then run the Completeness, Compliance, and Correctness jobs.

Completeness: Required vs. Recommended

To require a field only on part of the hierarchy, use a dictionary override. For example, Serial Number lives on cmdb_ci and isn't required there, so override it on Hardware to make it mandatory for Hardware and its children. If a mandatory field makes Discovery or other IRE inserts fail, set glide.required.attribute.enabled to false (Configuration > CMDB Properties > Identification/Reconciliation Properties).

The base system ships no predefined required or recommended CI fields. Non-discoverable fields like Support Group, Change Group, and Managed By Group belong in Recommended, not Required.

Compliance: audits

1. Certification Filter: the scope (for example, all UNIX servers in one data center).
2. Certification Template: the conditions to pass. It references the filter.
3. Audit: runs the comparison on a schedule and can create tasks. Its type must be Desired State.

• Set up all three from CI Class Manager > class > Health > Compliance, or from Compliance > Filters, Templates, and Audits.
• Only Desired State and Scripted audits count on the Compliance scorecard. Audit types such as Compliance and Compliance Architecture are for other products.
• Scripted audits define the pass criteria in a script when template conditions are too limited. They need the certification_admin role, can't be created in CI Class Manager (use Compliance > Scripted Audits > Audits), and the base system includes a sample, "Data Center Zone Usage."
• After an audit runs, activate and run the CMDB Health Dashboard - Compliance Score Calculation job (Configuration > Health Preferences > Scheduled Jobs). Run health jobs as an admin, or results can be inaccurate.
• Don't confuse the two: a Desired State audit checks automatically that a field is filled in (ServerXYZ has an owner). A Data Manager certification policy has a person confirm the value is right (Joe Smith really owns ServerXYZ).

Correctness: Duplicate, Orphan, Stale

Duplicates: how the IRE handles them and how you fix them

• glide.identification_engine.skip_duplicates (true) and ...skip_duplicates.threshold (5): when the IRE matches several duplicates under the threshold, it updates the oldest and tags the rest with Duplicate of. The IRE skips any CI whose Duplicate of is populated.
• Duplicate of replaced Discovery source as the duplicate marker, so Discovery source now keeps the last real source.
• De-duplication tasks are created automatically, with no configuration. Work them from the CMDB Workspace (Home, Important actions) or the De-duplication Tasks module.

Reclassification

During identification, the IRE can move a CI to a better class. It's automatic by default. If the relevant property is false, the CI isn't touched and a reclassification task is created for review.

Granular control

• glide.identification_engine.update_without_{upgrade|downgrade|switch}_enabled are false by default. Set to true, the IRE applies the attribute updates but skips the reclassification. They take precedence over the legacy properties, but they're global (all or nothing).
• Reclassification Restriction Rules (cmdb_ire_reclassification_restriction) block a downgrade or switch between specific source and target classes while still allowing attribute updates. Source/Target inheritance = true extends the rule to child classes.

Remediation: task, rule, workflow

• Staleness, Orphan, Required, and Recommended all follow that same setup. First define staleness rules, orphan rules, and required/recommended fields in CI Class Manager.
• Two exceptions don't follow the flow: compliance audit tasks (configured in CI Class Manager > Health > Compliance) and de-duplication tasks (automatic).
• Example from the course: a stale CI triggers a workflow that re-runs Discovery. If it succeeds, the CI is no longer stale. If not, the workflow sets Non-Operational/Retired and assigns to Hardware.
• Now on Now: automating CMDB audits saved about 1,252 hours a year, a 60% improvement.

Life cycle management and field sync

Effective life cycle management: understand each data type's life cycle (containers live days, servers live years), identify data owners with the Managed By Group, set up life cycle fields, then automate with Data Manager policies.

• Don't manage every status field. With legacy fields, pick Install Status or Hardware Status unless you have a business reason for both.
• Shared-field mappings live in Asset-CI Field Mappings and can be extended (for example, asset Comments → CI Short description). Baseline sync runs through business rules.
• Legacy sync is one-way in places: asset State → CI Install Status, but changing Install Status does not update asset State.

Click Enable Life Cycle Sync for a one-time sync from legacy to CSDM values, then resolve unmapped values in the Discrepancy Report. After that, legacy and CSDM fields stay in sync across the CI, the asset, and the related Install Base Item. End-of-Life statuses include Disposed and Donated.

CMDB Data Manager

A policy-driven framework for bulk end-of-life and certification work. The new Data Manager lives in the CMDB Workspace (legacy Configuration > CMDB Data Manager is deprecated by Yokohama). Old policies are compatible. Workspace 6.0.1 added certification policies. You create and publish policies from the Management tab; assignees work tasks from My Work.

2. Gain Insight20% of the exam

CMDB Query Builder

• Query types: CMDB Query, Service Mapping Query, and Combination Query (a saved Service Mapping query embedded in a CMDB query).
• Can mix CMDB and non-CMDB tables ("high-priority incidents for servers in Paris").
• Results can populate a CMDB Group, be exported, or be scheduled by email (daily, weekly, monthly, periodically, once, on demand).
• Move queries between instances with Export query (XML), then Import query from the Saved Queries tab.
• Interface: the CMDB Classes, Non-CMDB Tables, and Saved Service Queries lists to drag from; the Filters pane; the Canvas where you connect classes with relationships and AND/OR; the Properties tab (extra result columns) and Applied Filters tab; and Query Results.
• Its system properties are under Configuration > CMDB Properties > Query Builder Properties.

Unified Map

Combines Dependency View and Application Service maps into one view centered on a home node. Badges show incidents and changes, the Related items panel shows active incidents and planned changes, and you can filter (discovery source, CI type, levels) and save filters as presets. It's used for impact analysis in Incident, Problem, Change, and Event Management.

Intelligent Search (NLQ)

Plain-language, multi-table queries without knowing the data model. glide.cmdb.query.nlq.activated (on by default) integrates it with Query Builder, so multi-table results show View in Query Builder. It keeps your five most recent searches. Sample searches live in sn_cmdb_ws_nlq_sample_search, editable by sn_cmdb_admin.

Data Foundations Dashboards vs. Health Dashboard

• Two dashboards: CMDB (configuration quality and risk) and CSDM (framework adoption; it has Foundation, Crawl, Walk, Run tabs).
• They show nothing until the base scheduled jobs (names contain "Get Well") run. Calculations live in GetWell Script Includes. The metrics are a subset of HealthScan, which you request through your solution consultant.
• Install the CMDB and CSDM Data Foundations Dashboards app from System Applications > All Available Applications. Open them from Configuration > CMDB Data Foundations Dashboard or CSDM Data Foundations Dashboard, or from the CMDB Workspace Management tab.
• To fill them right away, open each "Get Well" job in System Definition > Scheduled Jobs and click Execute Now.
• Focus on Critical and High indicators first.

Value to other products

A well-configured CMDB can cut IT costs by up to 40%.

3. Ingest Data19% of the exam

Ingestion tools

Aim for about 90% of CMDB data populated automatically and 10% maintained manually under governance.

ServiceNow Discovery

• Phases in order: Scan → Classification → Identification → Exploration.
• Types of horizontal discovery: Network (finds internal IP networks), CI (devices, computers, and applications; the one run most often), Cloud (AWS and Azure resources), and Serverless (applications on hosts without a host scan first, using patterns).
• Manage schedules from the Discovery Admin Workspace. Open the Configuration Console from Discovery Definition > Configuration Console; excluding a CI type disables its probe or classifier.
• The MID Server sits inside your network. The instance sends probes; results come back to sensors. Protocols: WMI (Windows), SSH (Unix/Linux), SNMP (network gear, printers).
• Start small. Keep schedules to /16 or smaller (about 65,000 IPs), and use ranges that exclude the network and broadcast addresses. Use the Discovery Configuration Console to switch off CI types you don't manage.
• Discovered CIs get Discovery source = ServiceNow. A record first created by Asset Management shows SNAssetManagement until Discovery finds it.
• Non-discoverable data to add afterward: Assigned to, Owned by, Support group, Change group, Managed by group.

Application service types

Population method tables: discovered/manual = cmdb_ci_service_discovered, tags = cmdb_ci_service_by_tags, calculated = cmdb_ci_service_calculated, query-based = cmdb_ci_query_based_service. All extend cmdb_ci_service_auto.

Imports and the IRE

• IH-ETL flow: source data lands in import set staging tables → the Robust Transform Engine uses the ETL transform map to build IRE payloads → the IRE identifies and reconciles → data is written.
• IH-ETL terms: a CMDB Application (Name + Discovery Source, for example SCCM) can have many ETL transform maps; each map has one Data Source.
• SGC customization: a customized script transform creates a skipped change on upgrade; a changed query means its updates are skipped; customized mappings aren't supported. Service Graph Connector Central is a CMDB Workspace tab.
• SGC facts: categories include partner cloud providers, endpoint and IT asset management, monitoring and observability, security, software/server/network apps, operational technology (OT), and partner-built connectors. Each connector is reviewed for IRE use and CSDM conformance, and all of them use IH-ETL. Using an SGC that goes through the IRE needs ITOM Visibility licensing, except SCCM. Customizing a connector or its mappings makes the customer fully responsible for supporting it.
• IH-ETL guided setup (ETL Transform Map Assistant), in order: Basic Details (data source and import set) → Preview and Prepare Source Data (cleanse and transform) → Select CMDB Classes and Map Source Data (plus relationships) → Preview and Schedule (test run, roll back if needed, then schedule).
• The Integration Commons for CMDB Store app (sn_cmdb_int_util) adds the CMDB Integrations Dashboard, with performance and error metrics for every SGC and custom IH-ETL integration.
• Import sets into the CMDB: IH-ETL is the more modern option. If you use a transform map, add an onBefore script that calls CMDBTransformUtil so rows go through the IRE; when it does, the transform map's coalesce values are ignored. Plan first: which classes, which attributes, how records are identified, which target fields are mandatory, and what custom logic you need.
• Non-CMDB tables: the IRE and IH-ETL can load only the tables in glide.identification_engine.non_cmdb_tables (cmn_location, cmn_department, cmn_cost_center, cmn_building, sys_user, sys_user_group). Each gets exactly one identification rule, managed from list views, not CI Class Manager.

Populating group data

Build chain: CMDB Group → Dynamic CI Group → Technology Management Service → Technology Management Service Offering → relate the Dynamic CI Group to the offering with Contains::Contained by → set the group fields. The business rule CSDM - Sync Group Attributes does the sync.

Where each piece lives: Configuration > CMDB Groups, Configuration > Dynamic CI Groups, CSDM > Technology Management Service, and CSDM > Technical Service Offering (the offering references the service as its parent). Add the relationship from the offering's Related Items with the green plus. View the sync script in System Definition > Business Rules, Advanced tab.

4. Configure the CMDB15% of the exam

CI Class Manager

The central tree view for the class hierarchy (Configuration Item → Hardware → Computer → Server → Linux Server). Children inherit parent attributes.

• Principal Class limits which CIs appear in the Configuration Item field on Incident, Problem, and Change. Adding or removing needs sn_cmdb_admin + personalize_dictionary.
• New classes: avoid when possible. First install the CMDB CI Class Models Store app, then extend an existing class, and only create net-new as a last resort. Roles: admin, or sn_cmdb_admin + personalize_dictionary. Custom tables start with u_cmdb_ci and need identification and reconciliation rules (created or inherited).
• To retire a custom class, find a baseline class in CI Class Manager and move the CIs over with attributes mapped.

The IRE and identification rules

Every source calls the IRE's APIs; nothing writes straight to the CMDB. Goals: fewer duplicates, fewer unwanted updates, no attribute flapping. Its four functions are Identification, Reconciliation, De-duplication, and Reclassification. Logical CIs (Business Applications, Application Services, Services, Offerings) are created manually with wizards. Foundation data bypasses the IRE.

• Default rules: Hardware on cmdb_ci_hardware, Application on cmdb_ci_appl. View them in Configuration > Identification/Reconciliation > CI Identifiers or CI Class Manager.
• Independent CIs can be identified alone. Dependent CIs (like applications) need a relationship to another CI.
• Allow null attribute = false means every criterion attribute must have a value or the entry is skipped. That's a common cause of unexpected duplicates.
• The Hardware rule tries, in order: serial number with serial number type, then serial number, then name, then MAC address with name. A match updates the CI; no match creates one. Since Washington, product_instance_id is the first entry, currently used only for healthcare CIs.
• Software usually has no serial number, so applications are identified by the process they run.
• Good identifiers are unique and unchanging: serial number, configuration file path. Poor ones: IP address, version, and name for multi-instance apps. If Version is a criterion, an upgrade creates a new CI.
• Allow fallback to parent's rules: if a child rule doesn't match, try the parent's.
• Optional conditions work on lookup fields by default. For non-lookup fields, set glide.identification_engine.enable_identifier_optional_condition to true.

Reconciliation

• No reconciliation rules ship in the baseline. Create one per authorized source. They govern updates, not inserts. Once a higher-priority source updates a record, lower-priority sources can't overwrite those attributes.
• Lowest order number = highest priority.
• To block inserts from a source, use IRE Data Source Rules (none in baseline).
• Reconciliation runs after identification, only on updates; it isn't part of creating a CI. Well-built rules handle about 90% of deciding which source is trusted.
• Data Refresh Rules let a lower-priority source update after the higher-priority one has been silent for a set number of days.
• cmdb_datasource_last_update (Data Source Histories) records the last update per source, but only once reconciliation rules exist.

CMDB 360 / Multisource CMDB

Without it, values from losing sources are thrown away. With it, every source's values (winners and losers) are kept in cmdb_multisource_data, so you can compare sources, change rules and recompute without a new load, and revert a source's contributions. It only sees data that went through the IRE.

To turn it on: activate the ITOM Discovery License plugin (com.snc.itom.vis.license) from System Applications > All Available Applications, then set glide.identification_engine.multisource_enabled to true. Each cmdb_multisource_data record (added in Paris) stores common attributes as columns and every attribute in a JSON blob. Before Paris, only the last source per record was tracked, in Data Source Histories [cmdb_datasource_last_update].

View the data in CI Class Manager, the CMDB 360 Data Preview link, the Multisource Data related list, and the CMDB 360 view in CMDB Workspace (dashboard showing number of sources, CIs not reported by sources, attribute gaps, and conflicting values). The view refreshes daily through the Multisource Dashboard Analytics Population job.

Recompute is run from CI Class Manager when reconciliation rules change or a source turns out to be bad. On a CI, the CMDB 360 Data Preview link is under Related Links. The CMDB 360 Saved queries card in CMDB Workspace shows up to 20 queries, newest first.

Static vs. dynamic reconciliation rules

Dynamic beats static for the same attribute, and a rule on the CI's own class beats a parent-class rule.

5. CSDM Fundamentals11% of the exam

CSDM is a shared set of service definitions across ServiceNow products with prescriptive modeling guidance. It is not a product, SKU, process guide, set of reports, or code, and it doesn't fix past implementations automatically.

Seven domains (CSDM 5)

Service types and tables

Yokohama relabels: Technical Service → Technology Management Service, Technical Service Offering → Technology Management Service Offering, and cmdb_ci_service_auto → Service Instance.

Foundation → Crawl → Walk → Run → Fly

Other CSDM facts from the practice bank

• Incident forms use Service Offering and Application Service. Prioritize by Business Criticality on the offering.
• Model a platform (like ServiceNow itself) as a Platform Host with Platform Applications that reference it.
• Track regulatory compliance by linking compliance identifiers to CSDM objects.
• Governance teams: Foundation = Enterprise Architects, Data Stewards, Process Owners, Platform Owner. Design = Enterprise Architects, Platform Owner. Technical = Technology and Application Service Owners, Platform Owner. Portfolio = Service Owners, Platform Owner.

What's new in CSDM 5

• A new Ideation & Strategy domain; Software Bill of Materials (SBOM), from the Store as SBOM Core (sn_sbom_core); the System Component Model (with Software Component and Service Offering models) and Product Feature; Value Streams; Teams; new life cycle stages and statuses; the DevOps Change data model (Store); Service Instance and AI CI classes (model only); and the relabels Technology Management Service, Technology Management Service Offering, and Service Instance.
• Service Instance (formerly Application Service, cmdb_ci_service_auto) now has siblings: Data, Connection, Network, Facility, and Operational Process service instances. Application services can be created manually, or with Service Mapping by entry point, tags, or dynamic query.
• The Digital System Model has four layers: (Business) Process, Service (Logical), Functional, and Infrastructure.
• Business services and technology management services are each single-level, not a hierarchy, and both are operational CIs used for impact analysis in Incident, Problem, and Change. Business services are also used for change approvals.
• Teams is a related list on CIs that tracks group assignments by type (Change Group, Managed by Group, Support Group) without adding more fields. Setting the group field on the CI syncs the Teams list.

Key CSDM tables

• Common data elements are not CIs. Before loading them, ask: is there a trusted source, are there several, how often does it change, is it deep enough for your CIs, and who maintains it? The location table has Source ID, Location Type (Region, Country, State/Province, City, Site, Building/Structure, Floor, Room), and Managed by Group.
• Product models aren't CIs either. CIs point to them with the Model ID field. The ten base types: Application, System Component (version-specific; includes Software Component, Service Offering, Firmware, and AI System Product models), Service, Software (SAM entitlement only), Content, Contract, Facility, Hardware, Consumable, and Enterprise Good.
• Design & Planning tables (Business Capability, Business Application, Information Object) are used by Enterprise Architecture and are not targets of Incident, Problem, or Change.
• Ideation & Strategy: a Product Idea can be promoted into demand, project, epic, or story; a Strategic Plan holds the mission, vision, and values; Goals are qualitative and Targets measure them. Common persona: Product Owner.

CSDM life cycle stages

Life Cycle Stage (the phase) plus Life Cycle Stage Status (the step within it) together give an object's status. Added in Paris, provided by ServiceNow, and not editable; each object only shows the choices that fit it.

• A discovered CI defaults to Operational / In Use. Tangible End of Life statuses include Disposed, Donated, Sold, RMA, Vendor Credit, Buy-out, and Lease Return.
• Product Instance 2.0 (Xanadu) syncs stage and status across the CI, asset, and install base item, and disables the old legacy-status sync. Turning it off later is hard and can lose data, so test in non-production first.
• There's no rush to migrate, but some features (like CMDB Data Manager) require the life cycle mapping to be enabled. Check reports, scripts, business rules, and integrations that use the legacy fields first.

Migrating into CSDM

1. Back up your table data.
2. Map attributes: best practice, keep, refactor, or don't need.
3. Run the dependency script (the most important step). It finds the reports, business rules, and scripts tied to the old table. It doesn't move data.
4. Refactor attributes on the target tables.
5. Migrate the data: change the class, fix dependencies, reload, validate.

Properties and tables

The easiest points on the exam. Drill these last.

Off the top of your head

Answer each one before opening it.