DevOps Quick Reference: Environments, CI/CD, Testing & Governance

All-in-one DevOps reference for CTA scenarios. Environments, CI/CD pipelines, testing strategy, and governance — combined for rapid review. Tables over prose — this is your pre-board cheat sheet.

Sandbox Comparison Matrix

	Developer	Developer Pro	Partial Copy	Full Copy	Scratch Org
Metadata	Full	Full	Full	Full	Source-pushed
Data	None	None	Template sample	Full production	Scripted
Storage	200 MB	1 GB	5 GB	Same as prod	200 MB
Refresh	1 day	1 day	5 days	29 days	Recreate (ephemeral)
Max lifetime	Persistent	Persistent	Persistent	Persistent	30 days
Performance testing	No	No	No	Yes	No
Production data	No	No	Sampled	Yes (mask it!)	No
CI/CD use	Manual dev	Manual dev	Integration test	Staging/perf	Pipeline validation

Environment-to-Test Mapping

Environment	Tests That Run There	Data Needed
Scratch Org	Apex unit tests, LWC Jest, static analysis (PMD/ESLint)	Test data factory scripts
SIT (Partial Copy)	Integration tests, cross-team validation, E2E smoke	Sandbox template data
UAT (Partial Copy)	Business user acceptance tests	Representative (masked) data
Staging (Full Copy)	Performance tests, final regression, smoke tests	Production-scale (masked) data
Production	Post-deployment smoke tests only	Real data

CI/CD Pipeline — Step by Step

flowchart TD
    A["Developer commits\nto feature branch"] --> B["Pull Request created"]
    B --> C["CI Pipeline triggers"]
    C --> D["Static Analysis\n(PMD + ESLint)"]
    D --> E{Pass?}
    E -->|No| F["Block merge\nNotify developer"]
    E -->|Yes| G["Create scratch org"]
    G --> H["Deploy source"]
    H --> I["Run Apex tests\n(85%+ coverage)"]
    I --> J["Run LWC Jest tests"]
    J --> K{All pass?}
    K -->|No| F
    K -->|Yes| L["Code review\napproved"]
    L --> M["Merge to develop"]
    M --> N["Deploy to SIT"]
    N --> O["Integration tests"]
    O --> P["Deploy to UAT"]
    P --> Q["UAT sign-off"]
    Q --> R["Deploy to Staging"]
    R --> S["Smoke + perf tests"]
    S --> T["Deploy to Production"]

    style F fill:#e76f51,stroke:#c45a3f,color:#fff
    style T fill:#2d6a4f,stroke:#1b4332,color:#fff
    style K fill:#f4a261,stroke:#d4823e,color:#000

CI/CD Tool Comparison

Tool	Type	Strength	Cost	CTA When to Recommend
GitHub Actions	General CI/CD	Flexible, free for public repos	Free/Paid	Teams already on GitHub
Copado	SF-native	Built for SF, compliance, user stories	Expensive	Regulated industries, admin-friendly DevOps
Gearset	SF DevOps	Best diff/merge/compare tooling	Moderate	Teams needing metadata comparison
AutoRABIT	SF DevOps	Full suite with backup/restore	Expensive	Enterprise with backup requirements
Azure DevOps	General CI/CD	Enterprise integration, boards	Paid	Microsoft-centric enterprises
GitLab CI	General CI/CD	Integrated repo + CI, self-hosted	Free/Paid	Self-hosted requirement

CTA Tool Selection Logic

Don’t recommend a tool — recommend criteria. “We need CI/CD that supports scratch org creation, automated Apex testing, and integrates with our existing Git provider. Given the team uses GitHub, GitHub Actions with the sf CLI is the most cost-effective choice.” The board cares about your reasoning, not brand loyalty.

Deployment Mechanisms Compared

Mechanism	Rollback	Version History	CI/CD	Dependency Mgmt	CTA Verdict
Change Sets	None	None	No	None	Legacy — migrate away
Salesforce CLI	Via source control	Git history	Yes	Manual	Enterprise standard
Unlocked Packages	Install previous version	Package versions	Yes	Declared	Best for modular orgs
Managed Packages	Install previous version	Package versions	Yes	Declared	ISV only
DevOps Center	Limited	Basic tracking	Partial	None	Transitional tool

Unlocked Package Architecture

flowchart TD
    CORE["Core Package\n(Objects, Fields,\nPermission Sets)"] --> SALES["Sales Package\n(Flows, LWC, Apex)"]
    CORE --> SERVICE["Service Package\n(Flows, LWC, Apex)"]
    CORE --> INTEG["Integration Package\n(Named Creds,\nExt Services, Apex)"]

    style CORE fill:#1a535c,stroke:#0d3b44,color:#fff
    style SALES fill:#2d6a4f,stroke:#1b4332,color:#fff
    style SERVICE fill:#4ecdc4,stroke:#3ab5ad,color:#000
    style INTEG fill:#f4a261,stroke:#d4823e,color:#000

Feature Flags — Implementation Options

Mechanism	Toggle Speed	Scope	Deployment?	Best For
Custom Metadata Type	Deploy required	Org-wide	Yes (metadata)	Feature gates across environments
Custom Settings	Instant (DML)	Per user/profile	No (data)	Quick kill switches, user-level control
Custom Permissions	Permission Set assign	Per user	Yes (metadata)	User-level feature access

Testing Quick Reference

Test Data Strategies

Strategy	Description	When
Test Data Factory	Centralized `@isTest` utility class	Always — default approach
`@TestSetup`	Creates data once for all test methods	Multiple methods need same base data
Static Resources	CSV loaded as test data	Bulk tests with specific patterns
`SeeAllData=true`	Tests see real org data	Almost NEVER — only for specific platform features

Apex Testing Checklist

Test data factory pattern (no hardcoded data in tests)
Positive tests (happy path)
Negative tests (error handling, invalid data)
Bulk tests (200 records — trigger bulkification)
Boundary tests (0 records, null values, max values)
Test.startTest() / Test.stopTest() for governor limit reset
HttpCalloutMock for all external callouts
Meaningful assertions (not just “it didn’t crash”)
85%+ coverage target (not 75%)

Test Automation Tools

Tool	Layer	Automated?	CTA Notes
Apex Testing Framework	Unit + Integration	Fully	Foundation — always include
Jest (`@salesforce/sfdx-lwc-jest`)	Component	Fully	Required for any LWC
PMD	Static analysis (Apex)	Fully	Free, catches common issues
ESLint	Static analysis (JS)	Fully	Standard for LWC JavaScript
Provar	E2E (SF-native)	Automated	Understands SF DOM
Copado Robotic Testing	E2E (SF-native)	Automated	No-code test creation

Governance Cheat Sheet

RACI — Simplified CTA Version

Activity	CTA/EA	Dev Lead	Admin Lead	PM
Architecture decisions	A, R	C	C	I
Code standards	A	R	I	I
Declarative standards	A	C	R	I
Production deployments	C	R	R	A
Vendor evaluation	A, R	C	C	C
Release planning	C	R	R	A

R = Responsible, A = Accountable, C = Consulted, I = Informed

Change Classification

Type	Risk	Approval	Example
Standard	Low	Pre-approved	Field label, report
Normal	Medium	CAB review	New automation, integration
Emergency	High	Expedited + post-review	Production bug, security patch
Major	High	Full CAB + ARB	New cloud, data migration

Compliance — What to Mention at the Board

Framework	Key SF Requirement	CTA Must-Say
GDPR	Right to be forgotten, consent management	”Data deletion processes, anonymization scripts”
HIPAA	PHI encryption, audit logging	”Shield Encryption, Event Monitoring, data masking in sandboxes”
SOX	Audit trails, segregation of duties	”Field Audit Trail, approval process segregation”
PCI-DSS	Cardholder data protection	”Never store card data in SF — use payment gateway integration”

Compliance Overrides Everything

In regulated scenarios, compliance requirements override all architectural preferences. Do not recommend a technically superior solution that violates compliance. Document compliance as architectural constraints.

Reverse-Engineered Use Cases

Scenario 1: Insurance Company — Legacy Modernization

Situation: Insurance company has 15 developers, mix of change sets and manual deployments, no CI/CD, 3 production deployments per year. New CTO wants DevOps modernization.

What you’d do: Phased approach.

Phase 1 (Month 1-2): Introduce Git (GitHub), establish branching strategy (modified GitFlow), move all metadata to source control. Keep change sets temporarily for admin changes.
Phase 2 (Month 3-4): Set up GitHub Actions pipeline — scratch org validation on every PR. Automated Apex tests + PMD static analysis as merge gates.
Phase 3 (Month 5-6): Migrate from change sets to CLI deployments. Introduce unlocked packages (Core, Sales, Claims, Integration). Increase release cadence to monthly.

Governance: Establish ARB (monthly reviews). CAB for all production deployments. Hybrid CoE with central architecture team and BU delivery squads.

Scenario 2: Retail — High-Volume Seasonal Deployments

Situation: Retailer with Black Friday traffic spikes. Needs zero-downtime deployments, feature flags for seasonal promotions, and performance testing with 10M product records.

What you’d do: Feature flags via Custom Metadata Types for all promotional features — deploy code weeks early, activate flags on launch day. Full Copy sandbox for performance testing with production-scale data (10M records). Perf test every SOQL query and batch job.

Deployment: Phased by user group (pilot store first, then regional rollout). Kill switch via Custom Settings for instant disable if issues detected.

Testing: Automated regression suite runs on every PR. Load testing on Full Copy sandbox simulating Black Friday API volume. UAT with business team 2 weeks before promotion launch.

Scenario 3: Government Agency — FedRAMP Compliance

Situation: Federal agency implementing Service Cloud. FedRAMP compliance requires formal change management, audit trails, and government cloud hosting.

What you’d do: Waterfall-hybrid methodology (formal milestones with agile sprints within each phase) to satisfy audit requirements. Copado for CI/CD — its compliance features generate deployment artifacts needed for audits. Shield Event Monitoring for all admin actions.

Environment strategy: Government Cloud with dedicated infrastructure. Full Copy sandbox for staging with complete data masking. No production data in any other environment. Post-copy scripts disable all external integrations and mask all PII.

Governance: Formal CAB with documented approval for every change. ARB quarterly reviews. All architectural decisions recorded as ADRs.

DevOps Maturity Quick Assessment

Level	Characteristics	CTA Recommendation
Level 1: Ad Hoc	Change sets, manual testing, no source control	Introduce Git + CLI, basic CI
Level 2: Managed	Source control, some CI, manual deploys	Automate pipeline, add scratch orgs
Level 3: Defined	CI/CD pipeline, automated testing, sandboxes	Add unlocked packages, performance testing
Level 4: Measured	Metrics-driven, feature flags, fast rollback	Optimize release cadence, continuous deployment
Level 5: Optimized	Full automation, canary releases, self-healing	Maintain and innovate

Key Anti-Patterns

Anti-Pattern	Fix
Deploying to prod on Fridays	Tuesday/Thursday deployment windows (off-peak)
No rollback plan	Document rollback for every deployment
Full Copy sandbox with unmasked PII	`SandboxPostCopy` Apex for mandatory masking
One developer with all the knowledge	Cross-training, documented runbooks, RACI
Skipping Staging environment	Staging is mandatory for performance validation
No destructive change review	Destructive changes need code review + CAB approval
”75% coverage is enough”	85%+ with meaningful assertions

Deep Dive Links

Environment Strategy — Full sandbox analysis and topology patterns
CI/CD & Deployment — Pipeline architecture, branching, unlocked packages
Testing Strategy — Test pyramid, automation tools, UAT planning
Governance Model — ARB, CAB, CoE models, ADRs, compliance
Risk Management — Risk register, probability-impact matrix
Trade-Offs — Agile vs Waterfall, change sets vs CLI, and more
D6 Quick Reference Overview — Back to overview