npx remotion cloudrun services deploy
Cloud Run is in Alpha, which means APIs may change in any version and documentation is not yet finished. See the changelog to stay up to date with breaking changes.
npx remotion cloudrun services deploy
Creates a new service in your GCP project. If a service exists in the same region, with the same Remotion version, with the same amount of memory, disk space and timeout duration, the name of the already deployed service will be returned instead.
Example output
Validating Deployment of Cloud Run Service:
Remotion Version: 3.3.95
Memory Limit: 2Gi
CPU Limit: 1.0
Timeout: 300
Project Name: remotion-example
Region: us-east1
Deploying Cloud Run Service...
Cloud Run Deployed!
Service name: remotion--3-3-95--mem512mi--cpu2--t-1200
Version: 3.3.95
CPU Limit: 2
Memory Limit: 512MiTimeout: 1200sec
Region: us-east1
Service URL: https://remotion--3-3-95--mem512mi--cpu2--t-1200-1a2b3c4d5e-ue.a.run.app
GCP Console URL: https://console.cloud.google.com/run/detail/us-east1/remotion--3-3-95--mem512mi--cpu2--t-1200/logs
--region
The GCP region to select. The site that the service will be accessing should also be in this same region to minimise latency.
--memoryLimit
The upper bound on the amount of RAM that the Cloud Run service can consume. Default: 2 GB.
--cpuLimit
The maximum number of CPU cores that the Cloud Run service can use to process requests. Default: 1.0.
--minInstances
The minimum number of service instances to have available, regardless of requests. Default: 0.
Any running instances, even if they are not performing a render, will be billable in GCP. The default minimum number of instances is zero, which means that when no requests are made to your service, you are not billed.
--maxInstances
The maximum number of service instances that can be create by GCP in response to incoming requests. Default: 100.
--timeoutSeconds
Timeout of the Cloud Run service. Default: 300 seconds.
Not to be confused with the --timeout
flag when rendering which defines the timeout for delayRender()
.
--onlyAllocateCpuDuringRequestProcessing
v4.0.221
If this is set to true, cpu_idle
will be set to true
in the service manifest.
CPU alloction will be disabled while no request is being processed, which can lead to significant cost savings.
--quiet
, -q
Only logs the service name, and 'Authenticated access granted'.