AWS AppConfig
Faster application deployments with automated safeguards
What is AWS AppConfig?
AWS AppConfig allows teams to move safer and learn faster. AWS AppConfig helps you modify application behavior safely in production without redeploying code, using feature flags and dynamic configurations to control runtime settings in real time. This approach reduces release risk, accelerates deployment, and enables rapid response to issues by allowing gradual feature rollouts to specific user groups, with impact monitoring and confident expansion.
AWS AppConfig experimentation tools, which are based on 25+ years of experimentation best practices within Amazon, support A/B and multivariate testing on live traffic, letting you assign user segments to treatment groups, measure outcomes, and make data-driven decisions before full-scale implementation.
Benefits
By updating configuration and flags while your application is live on production, you can adjust software behavior immediately or gradually without the risk of redeploying code. AWS AppConfig empowers your teams to stay responsive—toggle features or throttle configurations in real time without restarting your production apps.
AWS AppConfig experimentation helps you safely experiment with front-end and back-end application changes using advanced audience exposure and feature flag–based treatments. Run A/B tests, multivariate tests, and gradual feature rollouts on real production traffic to measure the impact of new features, configurations, and user experiences before broad release. Built on Amazon’s 25+ years of internal experimentation practices, AWS AppConfig experimentation helps teams make data-driven decisions, reduce operational risk, and confidently scale successful changes across their applications.
AWS AppConfig supports best practices for safe software delivery by enabling configuration changes to be deployed instantly or gradually. You can monitor deployments using operational metrics and define a monitoring period to evaluate application health after a change is released. By integrating with alarms in Amazon CloudWatch, Datadog, Dynatrace, or other APMs, AWS AppConfig can automatically roll back configuration changes when predefined thresholds are breached, helping reduce operational risk and minimize the impact of failed deployments.
AWS AppConfig enables teams building AI-powered applications to manage LLM model settings, parameters, and prompts as dynamic configuration rather than hardcoded values. Update model versions, adjust temperature and token limits, or deploy new prompt variations without code changes or redeployments. Combined with built-in experimentation, teams can A/B test models and prompt variants on live traffic, and promote winners based on measured outcomes, turning AI experimentation into a data-driven optimization loop.
PBS
Thanks to AWS AppConfig, we have established a mechanism for safer deployment practices with less risk involved. As a result, we can save time and move faster. Our engineering team has been able to safely roll out changes that would have normally taken days in hours.
Use cases
Many customers use AWS AppConfig to safely release new features. A feature can be released on production, but hidden behind a feature flag, inaccessible to users. When the feature is ready to be released, deploying a small configuration update with AWS AppConfig can “turn on” that feature for users. Features can be released to some or all users, instantly or gradually with multiple values within flag data, allowing customers to target those values to fine-grained and high-cardinality user segments.
AWS AppConfig can be used to adjust ops behavior at runtime. For example, you may put timeout values into configuration data, and want to update those on the fly to best tune application performance. Similarly, a DevOps engineer may want to adjust logging verbosity, and increase what events are recorded during a suspected security incident, all without re-deploying the full app.
Certain functionality, APIs, or types of data may only be available to specific users for your app. This can be managed through a dynamic allow-list. Conversely, you may want to block access to specific data or features, and this would be managed through a block-list. AWS AppConfig enables quick and safe management of these lists that gate access.