How do we configure approval workflows and scheduled releases in LaunchDarkly?

Moving fast without losing control starts with two things: approvals and scheduling. In LaunchDarkly, you can require approvals before risky changes go live and schedule progressive rollouts so features land safely—after deploy, no redeploys required.

Quick Answer: LaunchDarkly lets you configure approval workflows per environment and tag, then automate when flag changes go live using scheduled changes and workflows. Together, they create a repeatable, auditable release pipeline that reduces blast radius and kills 2am fire drills.

The Quick Overview

• What It Is: A runtime control surface for approval workflows and scheduled releases built into LaunchDarkly feature flags and workflows.
• Who It Is For: Engineering, product, SRE, and platform teams that need governance (approvals, audit trails) and automation (scheduled rollouts, auto-rollbacks) for changes in production.
• Core Problem Solved: Shipping changes is either too slow (heavy change boards) or too risky (no guardrails). This gives you both speed and safety in the same release process.

How It Works

LaunchDarkly decouples deploy from release. Code ships behind flags; approvals and schedules control when and how those flags change in production.

At a high level:

1. Define governance: Decide which environments, flags, and segments require approval, and who can approve.
2. Wire approvals into your workflow: Use LaunchDarkly approvals or ServiceNow change requests, with optional auto-apply when external approvals complete.
3. Automate releases: Use workflows to schedule changes to targeting rules, percentages, and kill switches—optionally with progressive rollouts and automatic rollback based on metrics.

1. Configure approval requirements

Approvals answer: “Who needs to say yes before this change hits production?”

In LaunchDarkly you decide:

• Which environments require approvals

  • Common pattern:
    • production: approvals required
    • staging, dev: no approvals, or lighter rules
  • This keeps iteration fast in non-prod and controlled in prod.

• Whether to use LaunchDarkly or ServiceNow

  • LaunchDarkly approvals: Lightweight, in-product approvals for flags and segments.
  • ServiceNow approvals: Use ServiceNow as the approval system of record; LaunchDarkly can wait for the external change to be approved and then apply the changes.

• Which resources are governed

  • Apply approvals to:
    • All flags and segments in an environment, or
    • Only those with specific tags (for example, critical, payments, ai-config).
  • Tag-based rules give you a “critical flags only” safety layer without blocking every minor copy tweak.

• Automatic application of approved changes

  • When integrating with ServiceNow, you can choose whether:
    • LaunchDarkly automatically applies changes when the external change request is approved, or
    • A human manually applies the change after reviewing.

This configuration turns LaunchDarkly into a policy engine at runtime: changes to critical flags simply won’t take effect without the right approvals.

2. Use approvals in day-to-day releases

Once configured, approvals become part of the normal release flow:

• Propose a change:

  • A developer or PM edits a flag’s targeting, rollout percentages, or segment.
  • Instead of instantly applying, they submit for approval.

• Review & approve:

  • Approvers (often SRE, staff engineers, change managers) review:
    • The exact diff in targeting and rollout
    • The environment(s) affected
    • Related context: Jira ticket, incident references, release notes
  • They approve or reject directly in LaunchDarkly, or via ServiceNow if integrated.

• Apply with auditability:

  • Approved changes are applied in production—no redeploys required.
  • Every step (who requested, who approved, what changed, when) is captured in audit logs, forming a clear change history for incident reviews and compliance.

This gives you change control without forcing “big bang” deploys. You can still deploy continuously; you just decide which releases demand governance.

3. Schedule releases and rollouts

Approvals answer “who,” scheduling answers “when” and “how fast.”

LaunchDarkly workflows let you:

• Schedule flag changes for a future date/time

  • For example:
    • Turn on holiday_promo_banner at 08:00 local time tomorrow.
    • Disable a risky feature Friday 17:00 before a maintenance window.
  • All without touching your code or doing a deploy at that exact time.

• Orchestrate progressive rollouts

  • Define a rollout sequence as a workflow:
    1. Start at 1% of traffic.
    2. Increase to 10% if error rates stay below a threshold.
    3. Go to 50%, then 100%—with pauses between steps.
  • This dramatically reduces blast radius versus all-at-once releases.

• Automate based on external metrics

  • Use metrics from external tools (APM, logging, observability platforms) to drive behavior:
    • If error rate spikes or performance degrades, automatically pause the rollout or roll back to a safe configuration.
  • This is where the runtime control plane is most powerful: the platform reacts in <200ms worldwide, no recompile or redeploy.

• Combine with approvals

  • High-value pattern:
    • Require approval to start a workflow in production.
    • Once approved, the workflow automatically executes the entire rollout sequence.
  • Humans make the decision to proceed; automation handles timing and risk reduction.

Features & Benefits Breakdown

Core Feature	What It Does	Primary Benefit
Approval Workflows	Require approvals for changes to flags and segments per environment and tag.	Governance and safety for critical changes without slowing all work.
Scheduled Flag Changes & Workflows	Schedule targeting changes, rollouts, and kill switches for future times or multi-step sequences.	Predictable, hands-off releases that don’t require deploys or late-night pushes.
Metric-Driven Automation & Rollback	Automate pausing or rolling back based on external metrics and predefined thresholds.	Reduce blast radius, recover instantly, and avoid 2am fire drills.

Ideal Use Cases

• Best for regulated or high-risk production environments: Because you can require approvals for production and critical-tagged flags, enforce change management, and retain a full audit trail.
• Best for time-sensitive or global launches: Because you can schedule and progressively roll out features to specific segments or regions without coordinating manual deploys across time zones.

Limitations & Considerations

• Approvals don’t replace good tagging and ownership: You still need consistent tags and clear ownership of critical flags to make approval rules meaningful. Invest in tagging standards (payments, authentication, ai-config, performance-sensitive) early.
• Automation requires good signal: Metric-driven workflows are only as strong as the observability data feeding them. Ensure your error and performance metrics are reliable and integrated before enabling automated rollbacks.

Pricing & Plans

Approvals and workflow automation are part of LaunchDarkly’s broader runtime control platform. Specific access can vary by plan and contract, especially for advanced workflow automation and enterprise governance.

• Team / Growth-style plans: Best for product and engineering teams needing feature flags, basic governance, and simple scheduled releases.
• Enterprise-style plans: Best for organizations needing formal change management, ServiceNow integration, granular approvals, advanced workflows, audit logs, and custom roles across many teams.

For precise plan details, talk to LaunchDarkly directly or your account team.

Frequently Asked Questions

Do we have to use ServiceNow for approvals, or can we approve directly in LaunchDarkly?

Short Answer: You can do both; LaunchDarkly supports native approvals and ServiceNow integration.

Details:
If you already run change management in ServiceNow, LaunchDarkly can treat it as the approval source of truth. You decide whether to automatically apply changes in LaunchDarkly when a ServiceNow change is approved, or to keep a manual step. If you don’t use ServiceNow, LaunchDarkly’s built-in approvals let you manage requests and approvals for flag and segment changes entirely within the platform, scoped by environment and tags.

Can we schedule a production rollout that also requires approval?

Short Answer: Yes. You can require approvals and then schedule or automate the rollout as a workflow.

Details:
A common pattern is:

1. Configure approvals for production and relevant tags (for example, critical).
2. Create a workflow that defines the rollout path (for example, 1% → 10% → 50% → 100%).
3. Submit the workflow for approval.
4. Once approved, LaunchDarkly executes the rollout automatically at the specified times, with any metric-based pauses or rollbacks you’ve defined.

This gives you a single, consistent release pipeline: controlled entry (approvals) plus safe, automated execution (scheduling and workflows).

Summary

Approvals and scheduled releases in LaunchDarkly turn feature flags from simple toggles into a full runtime control system. You can:

• Require approvals in the environments and flags that matter most.
• Use tags to focus governance on high-risk changes, not every cosmetic tweak.
• Schedule and progressively roll out features without redeploys or late-night pushes.
• Automate pause and rollback based on real production metrics.

The result is a release process that moves at the speed of your deploy pipeline but with the safety of a change board—just with less drama and far fewer 2am alarms.

Next Step

Get Started