Platform
Resources
Release pipelines: Safer, staged rollouts across your infrastructure
Infra teams now have an additional guardrail—a last line of defense when changing configs.
At Statsig, we believe you can move fast without breaking things. That’s why we made feature flags a core tool in our platform—table stakes for modern software development.
Many of our customers, like Atlassian, OpenAI, and Notion, ship at lightning speed—deploying hundreds, if not thousands, of changes per week. At such massive scale and velocity, even a small percentage, like 0.01% odds of a risky feature getting past CI/CD pipelines and human checks, can become a significant risk in production.
To mitigate this risk, Statsig users typically roll out feature flags gradually, shipping to 1% of users and then progressively rolling out to the rest.
Today, we're taking this approach to the next level with our newest feature, release pipelines: a powerful new way to safely deploy changes across your entire infrastructure.
When feature flags aren't enough
Consider this scenario: Your team builds a new feature and decides to roll it out to just 1% of users to start. Chances are, everything will be fine...but what if that feature contains resource-intensive code that impacts server performance?
Your 1% of users could be distributed across hundreds of clusters, and if this change causes unexpected behavior in production, it could bring down your entire infrastructure stack, as every server experiences the increased CPU and memory usage.
Suddenly, the other 99% of users—who haven't even seen the new feature—experience degraded performance or complete outages.
The odds may be very low, but once you reach a certain release velocity and CPU footprint, even small risks can have a significant impact. Traditional feature flags help with user segmentation but don’t address infrastructure-level rollout challenges.
This is where our release pipelines come in.
What are Statsig release pipelines?
Release pipelines enable you to implement multi-stage rollout strategies that respect your infrastructure boundaries.
With release pipelines, you can:
Roll out changes environment by environment (dev → staging → prod)
Target specific infrastructure segments within environments (prod-us-west → prod-us-east → prod-eu)
Control progression between stages with time intervals or manual approvals
Monitor each stage before proceeding to the next
Roll back instantly if issues arise at any stage
The benefits of using release pipelines
The cost of an outage can be massive, both financially and in terms of user trust.
Release pipelines provide a safety net when making any change. This is especially valuable for platform teams managing complex, multi-environment, multi-region infrastructure.
By controlling and delaying propagation across your infrastructure, you can:
Catch issues early, before they affect a large portion of your infrastructure
Prevent cascading failures across your entire system, ensuring higher uptime
Validate changes in real production environments with minimal risk
Learn more about how you can get started in the release pipeline documentation.
How we’re different: Our dual user and infra-level approach
Most feature flagging tools focus solely on user segmentation, leaving your infrastructure vulnerable during rollouts.
That’s why Statsig covers both:
User-level control: Targeted rollouts to specific user segments
Infrastructure-level control: Progressive deployment across servers, regions, and environments
Our dual-layer approach gives engineering teams unmatched control over how changes are rolled out to both users and infrastructure. This also reduces developer overhead by automatically deploying changes to a small set of clusters first, giving infra teams more time to address issues before full propagation by default.
Available in beta today
Release pipelines are currently in beta and available for enterprise customers. Contact us to get this feature enabled for your organization.
Join the Slack community
