Scaling

Sprintsail supports both manual scaling and automatic horizontal scaling based on metrics.

Manual scaling

Instances

Scale the number of running containers:

ss scale my-app --instances 3

CPU and memory

Adjust compute resources per instance:

ss scale my-app --cpu 500m --memory 1Gi

Combine all options:

ss scale my-app --instances 4 --cpu 1000m --memory 2Gi

View current scale

ss app info my-app

Name:       my-app
Status:     running
Instances:  3/3 running
CPU:        500m per instance
Memory:     1Gi per instance
URL:        https://my-app.my-org.sprintsail.com

CPU and memory units

Unit	Meaning
250m	0.25 vCPU
500m	0.5 vCPU
1000m or 1	1 vCPU
128Mi	128 MiB RAM
512Mi	512 MiB RAM
1Gi	1 GiB RAM

Autoscaling

Enable horizontal pod autoscaling based on CPU or memory utilization:

ss autoscale my-app --min 2 --max 10 --cpu-target 70

This scales between 2 and 10 instances, adding capacity when average CPU exceeds 70%.

Autoscale on memory

ss autoscale my-app --min 2 --max 8 --memory-target 80

Autoscale on both

ss autoscale my-app --min 2 --max 10 --cpu-target 70 --memory-target 80

When multiple targets are set, scaling happens when any target is exceeded.

View autoscaling status

ss autoscale info my-app

Autoscaling:  enabled
Min:          2
Max:          10
CPU target:   70%  (current: 42%)
Memory target: 80% (current: 55%)
Instances:    3
Last scaled:  12 minutes ago (scale up: CPU 78%)

Disable autoscaling

ss autoscale disable my-app

This keeps the current instance count but stops automatic adjustments.

Configuration via ss.yaml

name: my-app
instances: 2
cpu: 500m
memory: 512Mi

autoscale:
  min: 2
  max: 10
  cpu_target: 70
  memory_target: 80

Scale-to-zero

Starter plan apps that receive no traffic for 30 minutes are automatically scaled to zero. The first request after scale-to-zero incurs a cold start (typically 2--5 seconds). To prevent scale-to-zero:

ss scale my-app --min 1

Growth and Enterprise plans do not scale to zero by default.

Plan limits

Each plan enforces maximum scaling limits per app:

Resource	Starter	Growth	Enterprise
Max instances	3	20	Unlimited
Max CPU per instance	500m	2000m	4000m
Max memory per instance	1Gi	4Gi	16Gi
Autoscaling	No	Yes	Yes
Scale-to-zero	Forced	Optional	Optional

Attempting to scale beyond your plan limits returns an error with a link to upgrade:

Error: Instance limit exceeded (3/3). Upgrade to Growth for up to 20 instances.
→ https://console.sprintsail.com/billing/upgrade

Scaling strategy

Guidelines for choosing instance size and count:

• More small instances work better for stateless HTTP services. They spread load and survive individual failures.
• Fewer large instances suit memory-heavy workloads like in-memory caches or ML inference.
• Start with 2 instances and autoscaling for production apps. A single instance has no redundancy during deploys or failures.
• Set CPU requests to match your steady-state usage. Autoscaling responds to spikes.

Next steps

• Monitoring --- track resource utilization and set alerts
• Plans --- detailed plan comparison
• Deploying --- rolling updates and deploy configuration