Release management in Unleash helps you standardize, automate and safeguard your feature rollouts. Instead of manually controlling each stage of a release, you can define a reusable, structured rollout process that progresses automatically while monitoring application health to prevent issues from affecting users.
Release management combines:
Release management structures your rollout process around milestones that execute sequentially. Each milestone defines what portion of users receives the feature using activation strategies. You control progression between milestones through manual triggers, time-based automation, or metric-based conditions.
For a step-by-step walkthrough, see Get started with release management.
The following concepts work together to enable release management:
Release templates standardize your rollout approach across features. Each template contains a sequence of milestones that represent stages in your release process, such as:

Each milestone contains at least one activation strategy that determines for whom the feature flag is enabled. You define the template once and apply it to multiple feature flags, ensuring consistency across your releases.

When you apply a release template to a feature flag in a specific environment, Unleash creates a release plan. The release plan contains the milestones from the template and manages the progression through each stage.

Automated progressions eliminate manual milestone transitions. You configure time intervals for each milestone, and Unleash automatically advances to the next stage when the interval elapses.
For example, you might configure a release to:
Automated progressions reduce the operational overhead of managing releases, especially for teams shipping features frequently or outside standard working hours.
Impact metrics connect application data to your release process. You define metrics in your application code using the Unleash SDK, such as error counts, request latency, or memory usage. Unleash stores and visualizes these metrics, allowing you to monitor feature health directly in the Admin UI.

Impact metrics support three types:
Safeguards use impact metrics to protect your release. You attach a safeguard to a flag’s environment, set a threshold on a metric, and choose what should happen when the threshold is crossed:
For example, you might configure a safeguard to pause the rollout if error counts exceed 10 in a 15-minute window, or to disable the environment entirely if latency spikes past a critical level.
See Configure safeguards for setup details.
Release management provides structure and safety for feature rollouts: