Progressive (or gradual) rollouts let you spread big or critical feature releases over a period of time by gradually rolling it out to an increasing audience over a period of time. You can customize the feature variation you want to roll out, the audience you want to target, as well as the interval and increment size of the ramp-up.
When setting up the exposure for your Progressive Delivery rule, you have the option to select one of two types of ramp-ups: linear and custom.
When configuring your linear ramp-up, you’ll have to define the following custom parameters:
- Start date of the rollout.
- Initial exposure: What percentage of the audience will immediately be exposed to the feature?
- Percentage increment step for rollout audience.
- Time interval between increments.
- Final exposure limit: Define the percentage at which you want the rollout to stop. Below this section, you will see a confirmation of when and at what exposure percentage the rollout will stop.
Use the “+Add step” button to add a custom increment. Each step you add requires selecting a date/time for that step to take place, and the resulting exposure rate.
Why use gradual rollouts?
Gradually rolling out your feature is a practical way of staying ahead of the quality of your releases and keeping track of infrastructure and product metrics by ramping it up over an extended period.
If you wish to release a new feature to your entire user base, it can make sense to use Progressive Delivery rules, for any of the following reasons:
- Make sure product metrics are performing well for a small sample size before ramping it up to a larger audience
- In case a bug is discovered, the feature can be rolled back quickly before majority of the user base sees it
- New features can sometimes put unforeseen load on your infrastructure, and gradual rollouts allow you to preemptively spot performance issues