• Formerly Platform.sh
  • Contact us
  • Docs
  • Login
Watch a demoFree trial

Scale applications predictably without re-architecting

Upsun lets teams adjust application capacity and performance characteristics without redesigning infrastructure or changing deployment workflows. Scaling is managed at the platform level, keeping applications responsive as demand changes.

Upsun is designed for teams that need to handle growth, traffic spikes, and variable workloads while maintaining stability and control.

What this enables

Scaling and performance on Upsun allow teams to adjust application capacity through the console, CLI, or API: 

  • Adjust compute and memory resources per application and service container
  • Add more application instances with a single command or click
  • Scale applications to handle traffic spikes and sustained growth
  • Maintain consistent performance across environments

How it works

Horizontal scaling

Applications can scale horizontally by adding or removing instances based on demand. This improves availability and resilience during traffic surges.

Autoscaling

Upsun supports autoscaling based on real resource usage, automatically adjusting the number of application instances when defined thresholds are reached. This helps teams handle traffic fluctuations without manual intervention. Autoscaling policies are configurable and can be tuned per environment to balance responsiveness, stability, and cost.

Vertical scaling

Each application runtime and service runs in its own container with allocated CPU, memory, and storage, including optional Guaranteed CPU profiles. Teams can adjust resources independently per environment to match workload needs, without rebuilding the application.

Performance-aware architecture

Upsun’s platform architecture is designed to support predictable performance. Resource allocation, isolation, and managed services reduce contention and variability. Teams can validate scaling behavior in preview or staging environments before applying it to production

What this solves

Overprovisioning and waste

Without flexible scaling, teams often allocate more resources than needed. Upsun allows fine-grained adjustments to reduce unused capacity.

Unplanned downtime during traffic spikes

Sudden increases in traffic can overwhelm static infrastructure. Platform-managed scaling helps applications remain available under load.

Complex scaling operations

Managing scaling across services and environments is operationally expensive. Upsun centralizes scaling logic and execution.

Key capabilities

  • Horizontal scaling for application runtimes
  • Independent vertical scaling for services and containers
  • Environment-specific resource configuration
  • Support for high-traffic and burst workloads
  • Consistent scaling behavior across environments
  • Management via Git, CLI, API, and Console

Enterprise considerations

Reliability and availability

Scaling improves fault tolerance by distributing load and reducing single points of failure.

Cost control

Resource allocation can be tuned per environment to balance performance and cost, especially for non-production workloads.

Operational simplicity

Scaling changes are managed through platform configuration, reducing the need for custom autoscaling systems or manual intervention.

How teams use this in practice

Seasonal traffic spikes

An e-commerce team scales application instances ahead of peak traffic periods to maintain performance during high demand.

Growth without disruption

A SaaS team increases resources incrementally as usage grows, without re-architecting or migrating platforms.

Environment-specific tuning

Platform engineers allocate smaller resource profiles for preview environments while reserving higher capacity for production.

Get started

Use Scaling and Performance to keep applications responsive as demand changes.

Start a free trialRequest a demo
UpsunFormerly Platform.sh

Join our monthly newsletter

Compliant and validated

ISO/IEC 27001SOC 2 Type 2PCI L1HIPAATX-RAMP
© 2026 Upsun. All rights reserved.