Build Pipeline
Render's build pipeline handles the tasks that occur before a new deploy of your service goes live. Depending on your service, these tasks might include:
- Running your build command (
yarn,pip install, etc.) - Running your pre-deploy command (for database migrations, asset uploads, etc.)
- Building an image from a Dockerfile
All pipeline tasks consume pipeline minutes. Each workspace receives an included monthly allotment of pipeline minutes, and you can purchase additional minutes as needed.
Professional workspaces and higher can enable the Performance pipeline tier to run pipeline tasks on larger compute instances.
View your current month's pipeline usage from your Billing page.
Pipeline tiers
Professional workspaces and higher can choose between two pipeline tiers: Starter and Performance.
Hobby workspaces always use the Starter tier.
| Tier | Specs | Description |
|---|---|---|
| Starter (default) | 2 CPU 8 GB RAM |
For Hobby workspaces, includes 500 pipeline minutes per month. For Professional workspaces and higher, includes 500 minutes per member per month (shared among all members). Recommended unless your pipeline tasks require additional memory or CPU. |
| Performance | 16 CPU 64 GB RAM |
Available only for Professional workspaces and higher. Runs tasks on compute instances with significantly higher memory and CPU. Does not provide an included monthly allotment of pipeline minutes. Performance pipeline minutes are billed at a higher rate than Starter minutes. Use this tier if your pipeline tasks require memory or CPU beyond what's provided by the Starter tier. |
Specs and pricing details for each tier are available from your Workspace Settings page in the Render Dashboard.
Setting your pipeline tier
Your pipeline tier is a workspace-wide setting. Every pipeline task across your workspace uses the same tier.
- In the Render Dashboard, go to your Workspace Settings page.
- In the Build Pipeline section, select a pipeline tier.
- Confirm your selection in the dialog that appears.
Pipeline minutes
While they're running, your builds and other pipeline tasks consume pipeline minutes. You can view your current month's usage from your Billing page.
Pipeline minutes are specific to their associated tier. You can't use Starter minutes with the Performance tier or vice versa.
Included minutes
Hobby workspaces receive 500 Starter-tier pipeline minutes per month. Professional workspaces and higher receive 500 Starter-tier minutes per member per month (shared among all members).
The Performance tier does not provide an included monthly allotment of pipeline minutes.
Running out of minutes
If you run out of pipeline minutes during a given month, you automatically purchase an additional allotment of minutes for your current tier, unless:
- You've reached your monthly spend limit, or
- You haven't added a payment method.
In the above cases, Render stops running pipeline tasks (including service builds!) for the remainder of the current month. You can reenable pipeline tasks by raising your spend limit (and adding a payment method if you haven't).
Setting a spend limit
You can set a maximum amount to spend on pipeline minutes each month. As long as you're under your limit for a given month, you automatically purchase an additional allotment of minutes whenever you run out.
- In the Render Dashboard, go to your Workspace Settings page.
- In the Build Pipeline section, click Set spend limit (or Edit if you're editing an existing limit).
- Specify a new limit in the dialog that appears.
Build limits
- Render cancels a build if any of the following occurs:
- Memory usage exceeds the limit for your pipeline tier.
- Disk space usage exceeds 16 GB.
- Your build command fails or times out (after 120 minutes).
- Your pre-deploy command fails or times out (after 30 minutes).
- Each Render service can have only one active build at a time.
- Whenever a new build is initiated, Render cancels any in-progress build for the same service.
- Builds don't have access to your running service instance's resources (such as memory or disk).
- This is because pipeline tasks run on completely separate compute.