ESC

AI-powered search across all blog posts and tools

35 Articles

Blog

Practical Salesforce articles for developers, admins, and architects.

Latest
What you'll learn
Agentforce agents are built from Topics (scope), Actions (what to do), and Guardrails (safety)
The Atlas Reasoning Engine plans multi-step actions before executing — not just pattern matching
Your existing Flows and @InvocableMethod Apex become agent actions with zero new code
14 min read · Mar 25, 2026Read →
What you'll learn
AI helps most with PR review latency, junior ramp-up, repetitive scaffolding, and test-class drudgery
AI should never own business logic decisions, security review, or schema design
Just turning everything on creates noise — overlapping AI tools without rules confuse the team
8 min read · Mar 25, 2026Read →
What you'll learn
Dynamic Forms is now generally available for all standard objects, eliminating page layout limitations
Flow now supports reactive screen components and improved collection handling with new operators
DevOps Center reaches GA with pipeline automation and conflict resolution tooling built in
9 min read · Mar 23, 2026Read →
Spotlight
F
Admin8 min advanced

Approval Process Architecture - Patterns That Scale

Beyond the basics: multi-level, parallel, and dynamic approval patterns that work in complex orgs.

Parallel approval steps let multiple approvers work simultaneously, cutting total approval time significantly
Dynamic approvers via formula fields or hierarchy eliminate hardcoded user IDs from approval processes
Combining Flow with Apex ProcessInstanceWorkitem actions gives you programmatic control without losing audit trails
F
ForcenautMar 21, 2026
What you'll learn
Dynamic Forms move field-level visibility control from page layouts into Lightning App Builder, enabling record-specific UI without multiple layouts
Dynamic Actions let you show or hide action buttons based on record field values, user profile, or device — declared entirely in App Builder
Both features require Lightning Experience and are not supported in all object types — check compatibility before planning a rollout
8 min read · Mar 19, 2026Read →
What you'll learn
Multi-org adds operational complexity that only pays off when you have genuine data sovereignty, acquisition, or compliance requirements
Salesforce Connect solves cross-org read access without replication, but has governor limit implications that must be planned for
Event-driven synchronization via Platform Events or Change Data Capture is the most resilient pattern for keeping data consistent across orgs
10 min read · Mar 17, 2026Read →
What you'll learn
Platform Events decouple the publisher from the subscriber, eliminating tight integration dependencies
Events are retained for 24 hours, enabling replay for disaster recovery and missed-message scenarios
Subscribers using CometD or Apex triggers process events outside the publishing transaction governor limits
9 min read · Mar 14, 2026Read →
More articles
What you'll learn
Request-Reply is the right pattern when you need a synchronous response to continue your business logic
Platform Events enable reliable Fire-and-Forget with at-least-once delivery guarantees and replay capability
Never choose an integration pattern based on familiarity — choose based on latency tolerance and failure requirements
10 min read · Mar 12, 2026Read →
What you'll learn
OWD sets the floor for access — every other mechanism can only open access wider, never restrict it further
Apex Managed Sharing persists independently of the standard sharing model and survives changes to other sharing rules
With-Sharing and Without-Sharing in Apex enforce or bypass the running user's record access — always declare it explicitly
11 min read · Mar 8, 2026Read →
More articles
What you'll learn
Source-driven development with SF CLI and scratch orgs eliminates the "it works in sandbox" problem by making environments reproducible
Automated deployment pipelines reduce deployment risk by running tests on every commit before anything reaches QA or production
Treat your sfdx-project.json and package directories as the single source of truth — never make metadata changes directly in sandbox
10 min read · Mar 5, 2026Read →
What you'll learn
The sf CLI (v2+) replaces sfdx — all commands use the sf prefix and shorter flag names
sf org open --path /lightning/n/YourApp lets you deep-link directly into any Salesforce page from the terminal
sf apex run test --synchronous blocks until results are available — ideal for CI pipelines
10 min read · Mar 1, 2026Read →
What you'll learn
API call limits are per 24-hour rolling window and scale with your user licenses — not just your edition
Bulk API 2.0 is the right choice for any operation over 2,000 records — it consumes API calls differently and is far more efficient
Monitor /limits endpoint proactively so you catch consumption spikes before they cause failures
10 min read · Feb 25, 2026Read →
Spotlight
F
Apex9 min intermediate

Batch Apex Patterns for Large Data Processing

Master the essential patterns for building robust, scalable Batch Apex jobs that handle millions of records without governor limit failures.

Use the Stateful interface only when you genuinely need to aggregate data across execute() chunks
Chain batches via Database.executeBatch() in the finish() method to build multi-stage pipelines
Always collect failed records in a List<Id> and log them to a custom object — never swallow errors silently
F
ForcenautFeb 20, 2026
What you'll learn
Custom indexes and skinny tables are the fastest wins for query performance on large objects
BigObjects offer unlimited storage for historical data without polluting your main org storage
Architecture decisions made at 10K records become critical at 10M — plan ahead
9 min read · Feb 20, 2026Read →
What you'll learn
Queueable Apex is the right choice when you need callouts, need to chain jobs sequentially, or need more than the standard 50,000 SOQL row limit per transaction
Transaction Finalizers (available since API 52.0) give you a reliable hook to detect job failure and implement retry logic
Chaining depth is limited to 1 level in test context — the 5-level cap applies to synchronous production context, while async context allows unlimited chaining
9 min read · Feb 16, 2026Read →
What you'll learn
WITH SECURITY_ENFORCED in SOQL throws a QueryException if the user lacks FLS access — use stripInaccessible for a gentler alternative
Shield Platform Encryption encrypts at rest at the database level — Native encryption only obscures, it does not protect against Salesforce admin access
Event Monitoring gives you login history, report exports, and API usage logs — essential for regulated industries
10 min read · Feb 11, 2026Read →
More articles
What you'll learn
Schedule-Triggered Flows and Scheduled Paths in Record-Triggered Flows solve fundamentally different problems and should not be confused
Scheduled Paths evaluate the wait condition at schedule time, not at the time the path was created — this causes common confusion around field value changes
Time zone handling in Scheduled Flows operates in UTC by default; always account for this when scheduling business-hours-relative automation
9 min read · Feb 6, 2026Read →
What you'll learn
Every Get Records element that runs before a screen is displayed adds a server round-trip — eliminate unnecessary ones with formulas
Reactive screens process component interactions client-side with no server call, dramatically improving perceived performance
Choice set efficiency and collection pre-loading are the most underused performance levers in complex screen flows
8 min read · Feb 2, 2026Read →
More articles
What you'll learn
Debug mode in Flow gives you step-by-step execution visibility without needing Apex logs
Collection filters let you avoid loops and formulas for subset operations
Subflows and dynamic choice sets dramatically reduce duplication across automations
8 min read · Jan 28, 2026Read →
What you'll learn
Prefer Flow for declarative logic that admins may need to maintain without a developer
Use Apex when you need complex logic, cross-object operations, or external callouts
Combining both is often the best architecture: Flow for orchestration, Apex for complexity
9 min read · Jan 25, 2026Read →
What you'll learn
Every DML element and action in a Flow should have a fault connector — unhandled faults surface as generic errors to users
Before-save flows (Record-Triggered, runs before save) cannot perform DML — use them for validation and field updates only
Store error details in a custom object from the fault path so you have an audit trail of flow failures
8 min read · Jan 21, 2026Read →
Spotlight
F
Flows9 min intermediate

Record-Triggered Flow Best Practices That Actually Matter

The patterns that separate production-ready flows from ones that bring down your org at 2am.

Before-save flows run synchronously in the save transaction — use them for field updates, never for DML on other objects
Entry conditions are your first line of defense against performance problems and runaway recursion
One flow per object is not always the rule — but uncontrolled proliferation will destroy your org
F
ForcenautJan 16, 2026
What you'll learn
LWC loads up to 3x faster than Aura due to native Web Components standard — no framework runtime to download
Aura is no longer receiving new feature investments; all new Salesforce platform capabilities target LWC
Migration is component-by-component, not all-or-nothing — Aura and LWC can coexist and communicate
9 min read · Jan 10, 2026Read →
What you'll learn
The wire service provisions data reactively — when its parameters change, it automatically re-fetches, eliminating lifecycle management boilerplate
Wire results are cached by the LDS cache and shared across components on the same page — this is a performance feature, not a bug
Use imperative Apex when you need explicit control over timing, need to call on user interaction, or need to handle mutually exclusive loading states
10 min read · Jan 6, 2026Read →
What you'll learn
The Trigger Handler pattern (using Domain and Service layers) eliminates logic in trigger files
The Unit of Work pattern consolidates all DML into a single bulk operation, protecting governor limits
The Selector pattern centralizes SOQL queries, making them reusable, testable, and easy to audit
11 min read · Dec 20, 2025Read →
More articles
What you'll learn
Custom Metadata Types are deployable and versionable — they travel with your package or changeset
Custom Settings (Hierarchy) are ideal for per-profile or per-user configuration overrides
Use Custom Metadata for anything that should be the same across all environments; use Custom Settings for environment-specific values
9 min read · Dec 12, 2025Read →
What you'll learn
Choosing lookup over master-detail when ownership and security should cascade is the most common and costly mistake
Polymorphic lookups to multiple object types create query complexity that compounds as the org grows
Schema decisions made early are expensive to change — design for scale before writing the first record
9 min read · Dec 12, 2025Read →
More articles
What you'll learn
Non-selective queries against large objects will hit governor limits under load even if they work in sandbox
Indexes only help when the filtered field has high selectivity — low-cardinality fields like Status can hurt more than help
SOQL inside loops is the single fastest way to blow the 100-query governor limit
10 min read · Dec 5, 2025Read →
What you'll learn
Coverage percentage measures lines hit, not logic verified — 100% coverage with zero assertions is meaningless
Test Data Factories and @TestSetup dramatically reduce setup boilerplate and improve maintainability
Always test negative paths, bulk scenarios, and async operations — not just the happy path
9 min read · Dec 1, 2025Read →
What you'll learn
One trigger per object is not just a best practice — it is the prerequisite for a maintainable trigger framework
A base TriggerHandler class with virtual context methods eliminates the boilerplate conditional logic that clutters every trigger file
Static bypass flags and recursion guards must be implemented at the framework level, not scattered across individual trigger classes
11 min read · Nov 24, 2025Read →
Spotlight
F
Architecture9 min intermediate

Governor Limits Every Salesforce Architect Must Know

A complete reference to Salesforce governor limits with exact numbers, visual comparisons, and bulkification patterns that prevent hitting them.

All governor limits reset per transaction, not per trigger or method call
Bulkification is not optional — it is the single most important Apex pattern
CPU time limit (10 seconds synchronous) is the hardest limit to diagnose and recover from
F
ForcenautNov 18, 2025
What you'll learn
System validations and before-save flows fire before your Apex triggers
Validation rules run AFTER before triggers, meaning triggers can manipulate data before validation
After triggers fire after the record is saved to the database but before the transaction commits
10 min read · Nov 11, 2025Read →
ESC
Press ESC to close