/learn/enterprise-architecture

Enterprise Architecture

Move from feature tour to operating design.

50 min3 lessons3 retrieval drills

The promise

What you will be able to do by the end of this module

Survive a security architecture review by treating API scope as a security decision and being straight about what the vendor does not publish.

  • Treat OAuth scope as a security decision and explain Management scope as arbitrary code execution, not 'write access'.
  • Restrict a server-to-server integration to the Client Credentials grant and justify why.
  • Answer 'what are your rate limits?' correctly — they are not published — and propose measuring them in the tenant.
  • Design custom fields against their real caps: 10,000 characters for text, 200 for URL and encrypted, signed 32-bit for numeric.
  • Survive an architecture review by raising blast radius and failure behaviour before the architect does.
Start from what you already know

At enterprise scale you are not selling features, you are answering an architecture review. The questions are about blast radius, integration seams, and what happens when something fails.

01

Discovery

Map the operating environment before the product

Enterprise endpoint management spans ownership, identity, networks, security, service management, change, data, and regional operations. The SE's job is to expose dependencies and turn them into a testable design.

  • Estate: device types, OSs, locations, ownership, network constraints.
  • People: admins, helpdesk, security, app owners, change approvers, auditors.
  • Process: onboarding, patching, incident, request, recovery, retirement.
  • Systems: identity, ITSM, security, backup, CMDB/ITAM, BI, APIs.
02

Architecture

Define authority and data flow

For each integration, document the trigger, data, direction, system of record, identity, error handling, and audit owner. Connector lists change; architecture principles endure.

  • Who can issue a device action?
  • Which system owns a ticket, asset, user, and security event?
  • How are secrets, roles, and service accounts governed?
  • What happens when an API or downstream service is unavailable?
03

Adoption

Design the rollout as part of the solution

A strong architecture includes migration waves, pilot scope, policy rings, technician enablement, operational ownership, success metrics, and a handoff into onboarding and Customer Success.

Field principle

A sound technical motion connects architecture, proof, implementation alignment, reusable assets, and Customer Success partnership.

Mechanics

How it actually behaves

Each item is tagged with how far it can be trusted. Verified means checked against official documentation; field means it is our recommendation, not a vendor claim; unpublished means NinjaOne does not state it at all.

Three OAuth scopes with very different blast radius

verified

Monitoring is read-only. Management can create and modify organizations and devices and run scripts. Control enables remote access. Almost nothing legitimately needs Management and Control together.

Management scope is arbitrary code execution

field

'Run scripts' via API means a Management-scoped client secret is functionally a domain-admin-equivalent credential. Say this to the security architect before they work it out themselves — it is the fastest credibility you will earn in that room.

Client Credentials is the server-to-server grant

verified

Restrict the application to that grant so an interactive flow cannot be used. Authorization Code and Implicit are also supported; a refresh-token scope is requested separately.

Only system administrators can configure OAuth tokens

verified

A practical constraint on delegating integration work, and a question that comes up in every RBAC review.

Rate limits are not published

unpublished

NinjaOne does not publish per-endpoint rate limits, page-size caps, or regional base URLs on the open web; the interactive reference sits behind a tenant login. Integrators consistently report per-endpoint throttling requiring batching and caching.

What NinjaOne does not publish

Per-endpoint rate limits, pagination caps, and regional API base URLs are not publicly documented. Measure them in the customer's tenant and write down what you observe rather than quoting a figure.

Discovery

Ask these, then listen

Who reviews a new SaaS control plane here, and what did they reject last time?

Listen for: The last rejection tells you the actual bar you must clear.

What is your system of record for assets, and does it agree with reality?

Listen for: Determines integration direction and who is authoritative.

What happens to your integration when this API is unavailable?

Listen for: Nobody plans for it. Raising it makes you the adult in the room.

Demo path

Three moves, in this order

  1. 1

    A Monitoring-scoped token doing reporting.

    Read-only is enough for most integrations. Start here.

  2. 2

    A custom field written by script and read as a condition.

    Business state becomes monitorable state.

  3. 3

    Scope separation across two applications.

    One integration, one scope. Never one credential for everything.

Objection handling

What they say, what you say, how you prove it

What are your API rate limits?

They are not published per endpoint, and I am not going to invent a number. Integrators report per-endpoint throttling that needs batching and caching. We should measure it in your tenant during the POC.

Proof method: Refusing to guess is the correct answer to an architect, and it is memorable.

We need every integration to have full access.

Management scope can run scripts on endpoints. If your reporting integration is compromised with that scope, the blast radius is your entire estate. Read-only where possible.

Proof method: Framing scope as blast radius rather than permissions moves the conversation to their language.

Where SEs blow it

Do not say these

  • Never invent a rate limit. An architect will test the claim.
  • Do not describe Management scope as 'write access'. It is code execution.
  • Do not design custom fields without checking the type caps first.

From NinjaOne's channel

Watch it explained

Short clips from NinjaOne's own channel that reinforce the mechanics above. Each one says why it is here and what it backs up. These are supporting context, not product demonstrations — the sourced claims stay in the mechanics section.

Why this is here: Blast-radius thinking, which is exactly how to present OAuth scope selection to a security architect.Watch on YouTube ↗
Why this is here: Credential separation — the same instinct behind never combining Management and Control scope.Watch on YouTube ↗
Why this is here: Script plus custom field plus condition, which is the integration seam most evaluations miss.Watch on YouTube ↗

Retrieval practice

Rapid fire

Answer aloud before opening each response.

01What is the first integration question?

Which system is authoritative for this object or workflow?

02What makes a rollout architectural?

It defines rings, delegation, failure handling, ownership, evidence, and adoption - not just agent installation.

03What should a POC hand off?

A validated design, gaps/risks, success evidence, rollout plan, and named owners.