ServiceNow CIS-Data Foundations

SERVICENOW · CIS – Data Foundations (CMDB and CSDM) (CIS-DF)

QH Verified

By Mr Sparkles

Log in to rate

Cheat Sheet

ServiceNow CIS-DF Exam Mastery

Data Foundations · CMDB & CSDM

Please use my guide to finalize your training for the ServiceNow CIS-DF exam. Reproduction of this is not allowed.

Created by:Mr. Sparkles

The Golden Rules

  • Govern is 35%. If you only drilled Class Manager screenshots, you are underprepared.
  • Identification answers "same CI or new?" Reconciliation answers "which source owns this field?" Fix ID first.
  • IRE runs on write / ingest paths. Opening Unified Map or running Health jobs does not invoke IRE.
  • Prefer Service Graph Connectors over custom Import Sets. Direct GlideRecord into cmdb_ci is how you manufacture duplicates.
  • Completeness ≠ Compliance ≠ Correctness. Mix those three and you miss easy points.
  • Business Application ≠ Application Service. Design-time product vs running instance. Crawl before you Fly.
1

Blueprint Weights ~60 Q / ~130 min

Domain Weight
Govern 35%
Insight 20%
Ingest 19%
Configuration 15%
CSDM Fundamentals 11%
Reality check: CIS-DF is a hard prereq for several CIS certs (ITSM, Discovery, HAM, SAM, SM, SIR, VR — confirm live list). Weights shift; check ServiceNow University before you book.
2

Health scorecard Govern 35%

Pillar What it measures
Completeness Required + recommended attributes populated
Correctness Duplicates, orphans, staleness
Compliance Desired-state / standards match
Six-metric framing (same story, finer)
Required Recommended Duplicate Orphan Stale Compliance
  • Default staleness often cited as ~60 days with no update — confirm in your instance rules.
  • Orphan rules often ship unconfigured. Build the relationship expectations; do not assume OOB magic.
  • Class Manager Completeness tab ↔ Completeness KPI. Correctness tab ↔ stale / orphan rules.
3

Govern toolkit study hard

Tool Job
CMDB Health Dashboard Ongoing quality KPIs / scorecards
Data Foundations Dashboard Structural fitness, best-practice gaps, IRE-bypass patterns + Playbooks
CMDB Workspace Ops cockpit: search, health, 360, map, remediation
Data Manager Lifecycle policies at scale (attest / remediate / retire / archive / delete)
Deduplication Wizard Pick master, merge fields, re-parent tasks / relationships
Principal classes Filter ITSM CI lookups — does not stop Health audits of other classes

Data Manager actions

Action Meaning
Attestation Owner confirms before action
Remediation Normalize / fill gaps at scale
Retire Status / lifecycle change; still a CI
Archive Read-only move out of active
Delete Permanent — last resort
Trap: Data Manager does not merge duplicates. That is the Deduplication Wizard. Sequence retire → wait → archive/delete when the stem asks for lifecycle hygiene.
Stewardship: Data Steward / Librarian remediates health findings. Architect / sponsor charters the program. Do not assign daily cleanup to the CAB chair.
4

IRE pipeline Config + Ingest

  1. Identification — match existing CI or insert. First definitive identifier entry can stop the walk.
  2. Reconciliation — which data source may update which attributes (precedence).
  3. Ambiguous multi-match → reject / de-duplication task. Do not silent-merge.
Need Use
Stable match key Serial number (prefer over volatile IP)
Child needs parent context Dependent identification
Junk values like "unknown" Inclusion / exclusion filters on identifiers
Protect steward fields Attribute-level recon (block blank overwrite)
Transform into CMDB CMDBTransformUtil (IRE on) — coalesce becomes redundant for matching
Scripted ingest IdentificationEngineScriptableApi
IRE runs: Discovery writes, IRE-enabled transforms, IdentificationEngine APIs. IRE does not run: Unified Map views, Health score jobs, casual Query Builder browsing.
5

CMDB 360 / Multisource

  • Keeps per-source field history so you can see Discovery vs SCCM vs manual without last-writer amnesia.
  • Enable via identification-engine multisource property (instance-specific name — know the concept).
  • Raw store commonly referenced as cmdb_multisource_data.
UI idea Job
Source Cards Which source owns this field value
Lineage How values flowed over time
Coverage Where sources agree / disagree
Conflict resolution Choose / revert with eyes open
Trap: deny_class excludes a class from 360 processing. It does not delete CIs from the CMDB.
6

CI Class Manager

Concept Remember
Job Governed CMDB schema UI (hierarchy, attrs, ID / recon rules)
Place attributes Lowest common ancestor (servers → cmdb_ci_server, not root cmdb_ci)
Extend Prefer extend OOB parents + u_ fields
Abstract No direct inserts — force leaf classes
Extensible May create subclasses — ≠ principal
Principal CI class Basic Info — shapes ITSM lookups
Upgradeability: Map to OOB attributes first. Twenty custom columns on root cmdb_ci is how you buy years of pain.
7

Ingest picker 19%

Need Prefer
Certified app feed exists Service Graph Connector
Infra auto-population Discovery (+ MID)
One-time file Import Set + IRE transform
Scheduled DB pull JDBC + scheduled import + IRE
Real-time / custom REST + IdentificationEngine API
SaaS / contracts / BA data Manual or import — not "more Discovery"
Sources fight on a field Reconciliation rule
  • MID Server polls outbound to the instance. It is not your load balancer and not a inbound hole you open for fun.
  • Keep field maps even when IRE replaces coalesce for matching.
8

Relationships & Asset ↔ CI

Type vibe Meaning
Depends on Functional need / dependency
Runs on Execution host
Contains / Hosted on Containment / hosting
Asset vs CI status (do not mash)
  • Asset State → often syncs toward Hardware Status
  • Operational Status = runtime posture
  • Install Status = classic install lifecycle
  • Retiring an asset does not mean "delete the CI and hope."
Trap: Depends on ≠ Runs on. Wrong relationship type breaks impact analysis even when both CIs exist.
9

Insight tools 20%

Tool Use when
Query Builder Multi-hop, relationship-aware CI queries
Unified Map Topology / dependency visualization
NLQ Plain language → structured query (tune synonyms)
Saved / 360 queries Reusable multisource comparisons
Playbooks Guided Foundation remediation > cowboy scripts
  • Depth = relationship hops. It is not your record limit.
  • Downstream from an app usually walks toward infrastructure it depends on.
  • Optional / left-join style nodes keep parents when a leaf is missing.
Trap: Query Builder does not schedule Discovery. Unified Map does not run IRE.
10

CSDM map 11%

Domain Think
Foundation Users, groups, locations, companies, cost centers
Design Blueprints — Business Application, capabilities
Build Services coming into existence
Manage Technical Services Deployed Application Services + tech stack
Sell / Consume Business Services + Service Offerings (SLA packaging)
Object One-liner
Business Capability What the business can do
Business Application Owned software product (design-time)
Application Service Deployed / running instance
Service Offering Sellable tier / commitment
Crawl → Walk → Run → Fly Infra → App Service → BA → Capability
Crawl first: Business Application ↔ Application Service mapping. Skipping Walk for portfolio theater leaves hollow services and useless impact analysis.
11

Health triage picker

Symptom Fix lens
Required / recommended field empty Completeness
Fails desired-state / standard Compliance
Two records are the same thing Correctness — duplicate (wizard)
Missing expected relationships Correctness — orphan (rules + edges)
Not updated in ages Correctness — stale (+ Data Manager lifecycle)
No identification rules / wrong class bucket Foundations Dashboard + playbooks
!

Common traps

Trap Reality
ID vs recon swapped Match first, precedence second
Health / Map "runs IRE" Only ingest write paths
Coalesce still required with IRE Matching moves to identification rules
Archive = Delete Archive keeps; Delete is gone
BA = Application Service Product vs running instance
Principal class = delete components It filters lookups
More Discovery for SaaS contracts Manual / import + recon protection
Study only Configuration Govern is over a third

Last-minute checklist

  1. Govern: Health 3Cs + six metrics, Data Manager vs Dedup Wizard, Workspace, Foundations Dashboard + Playbooks
  2. IRE: Identify → Reconcile; serial > IP; dependent ID; multi-match rejects
  3. Ingest: Connector > custom; CMDBTransformUtil; protect non-discoverable attrs
  4. Class Manager: LCA attributes, extend OOB + u_, principal ≠ extensible
  5. 360: per-source history; deny_class ≠ delete
  6. Depends on ≠ Runs on; Asset State ≠ Operational Status
  7. Insight: Query Builder depth ≠ limit; Map visualizes only
  8. CSDM: BA ≠ App Service; Crawl before Fly; compliance IDs on stable logical objects