CTA Awareness flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

CTA review board

A live scenario defense in which the candidate presents a solution design, justifies trade-offs, and answers probing questions from a panel of experienced architects.

Non-functional requirements

Constraints such as volume, latency, availability, retention, and compliance that shape architecture as strongly as functional features do, and must be elicited explicitly during discovery.

Declarative-first principle

Standard functionality and declarative configuration are chosen before custom code because they cost less to maintain and carry lower upgrade risk on a shared multi-tenant platform.

Master-detail relationship

A relationship where the child record cannot exist without the parent, inherits ownership and sharing, is cascade-deleted, and supports roll-up summary fields on the parent.

Junction object

An object with two master-detail relationships that resolves a many-to-many association, for example linking contacts to multiple custom campaign records with per-link attributes.

External ID field

A field flagged as an external identifier that enables upsert matching, data loader correlation, and stable references between the platform and outside systems.

Data skew

Uneven distribution of records across owners or parent records, which causes record lock contention and slow sharing recalculation during high-volume inserts and updates.

Bulk API 2.0

An asynchronous, job-based REST interface designed for very large record volumes, using separate batch limits and suited to migrations and nightly synchronization rather than interactive use.

Idempotent inbound integration

A design where repeated delivery of the same request produces no duplicate records, typically achieved with external ID upserts or a deduplication key checked before insert.

Platform Events

A publish-subscribe mechanism carrying defined event messages between systems, supporting replay identifiers so subscribers can recover from missed or failed deliveries.

Composite API

A REST capability that bundles multiple subrequests into one call, reducing round trips and letting later subrequests reference records created earlier in the same payload.

Organization-wide defaults

The baseline record access level for an object that sets the most restrictive sharing posture, which role hierarchy, sharing rules, and manual shares can widen but not tighten.

Permission set groups

A container that bundles permission sets into a single assignable unit, letting architects model job functions additively instead of maintaining many cloned profiles.

JWT bearer OAuth flow

A server-to-server authentication flow where a signed assertion replaces user interaction, commonly used for integration users that run without a human present.

Field-level security in Apex

Field permissions hide fields in the interface but do not block SOQL in code, so Apex must enforce them using WITH SECURITY_ENFORCED or stripInaccessible explicitly.

Shield Platform Encryption

Encryption at rest with customer-managed keys; deterministic schemes preserve exact-match filtering while probabilistic schemes prevent filtering and sorting on the encrypted field.

Governor limits

Per-transaction caps on CPU time, queries, DML statements, and callouts that protect the shared multi-tenant infrastructure and force bulkified, asynchronous designs at scale.

Queueable Apex

An asynchronous job type that accepts complex parameter types, returns a job identifier, and can chain follow-up jobs, making it preferable to future methods for most async work.

Apex sharing keywords

The with sharing, without sharing, and inherited sharing declarations that determine whether the current user's record access rules are enforced during code execution.

Custom metadata types

Configuration records that deploy between environments like metadata, are available in formulas and validation, and are preferred over custom settings for application configuration.

Query selectivity

A filter condition is selective when it matches a small fraction of rows and uses an indexed field, allowing the query optimizer to avoid a full table scan.

Skinny table

A support-created narrow copy of frequently filtered fields that speeds large-volume queries, but must be refreshed whenever the source object schema changes.

UNABLE_TO_LOCK_ROW

A runtime error caused by concurrent transactions contending for the same parent record or skewed data, usually mitigated by retry logic and distributing ownership.

Batch Apex scope size

The number of records processed per batch chunk, defaulting to 200; tuning it trades heap and CPU per chunk against total batch count and elapsed time.

Partial Copy sandbox

A sandbox that includes metadata plus a sample of production data defined by a template, useful for testing with realistic but reduced data volumes.

Apex test coverage requirement

Production deployment requires at least seventy-five percent overall coverage and all tests passing, though coverage alone does not demonstrate meaningful assertion quality.

Architecture decision record

A short document capturing the context, the decision made, and its consequences, preserving rationale so future architects understand why a design path was chosen.

Build versus buy

A trade-off comparing custom development against standard features, AppExchange products, or external systems, weighed on fit, cost, risk, and time to value.

Risk register

A structured list pairing each identified risk with likelihood, impact, an accountable owner, and a mitigation, reviewed as part of solution design approval.

Phased delivery

Sequencing work so the data model and security model are proven first, followed by integration and then scale optimization, reducing the cost of late discovery.

Trust boundary in diagrams

A marked line where identity is asserted or data crosses between security zones, showing reviewers exactly where authentication and authorization must be validated.

Audience-tailored narrative

Presenting the same architecture differently for executives, who need cost, risk, and timeline, and engineers, who need data model, limits, and failure modes.