Replicas & Scaling

OneLift provides a complete scaling suite: manual scaling, automatic scaling, resource limits, and capacity monitoring.

Manual Scaling

Set the replicas field in .lift.json or use the CLI:

{
  "replicas": 3
}

# Scale to 3 replicas
lift scale 3

# Check capacity before scaling
lift capacity

Containers are named <app>-1, <app>-2, <app>-3 and load-balanced by Traefik.

Load Balancing & Sticky Sessions

Traefik distributes traffic across replicas using round-robin. Each new visitor (without a session cookie) may land on a different replica.

Once a visitor connects, a sticky session cookie (onelift_instance) is set. All subsequent requests from the same browser go to the same replica. This prevents session loss for stateful applications like n8n, Strapi, or any app that stores sessions in memory.

• Same user (same browser/cookie) → always hits the same replica
• Different users → distributed across replicas
• New incognito window → may hit a different replica

Autoheal (Automatic Recovery)

OneLift runs a server-level autoheal container that monitors Docker healthchecks and automatically restarts unhealthy containers across all tools.

• Runs every 30 seconds
• Restarts containers that fail their Docker healthcheck
• Container restarts are logged to the Activity page
• Managed automatically — no user configuration needed

Check autoheal status with:

lift self-status

Automatic Scaling

Configure CPU-based autoscaling with predefined profiles:

# Use a profile
lift autoscale strapi --profile balanced --min 1 --max 5

# Custom CPU target
lift autoscale strapi --cpu-target 75 --min 2 --max 8

# Disable
lift autoscale strapi --off

The scaler daemon monitors CPU usage via docker stats and adjusts replicas within the defined bounds.

Resource Limits

Set memory and CPU limits per tool:

# View current limits
lift limit

# Set limits
lift limit --tool strapi --memory 2g --cpu 2

Limits are applied immediately via docker update without container restart.

Capacity Planning

Check server resources and per-tool headroom:

lift capacity

Shows CPU, RAM, disk usage and calculates how many more instances of each tool can fit.

Real-time Monitoring

When the Docker Monitoring agent is installed, the Scaling panel shows live CPU, RAM, and uptime for each container. Data refreshes at the agent's configured interval (default: 60 seconds).

Without the agent, metrics are only fetched on page load (one-time snapshot via docker stats).

To install the agent, go to Agents → Overview and enable Docker Monitoring.

Infrastructure Management

lift self-status    # Check Traefik, Autoheal status
lift self-logs      # View deploy/operation logs
lift self-restart   # Restart infrastructure services
lift self-rollback  # Rollback CLI to previous version

Scaling Marketplace Tools

Marketplace tools (PostgreSQL, Redis, n8n, Grafana, etc.) support the same scaling commands via the --path flag, which points to the tool's install directory under /opt/onelift/tools/.

# Scale PostgreSQL to 2 replicas
lift scale postgresql 2 --path /opt/onelift/tools/postgresql

# Scale a specific service in a multi-service tool
lift scale strapi 3 --path /opt/onelift/tools/strapi --service app

# Check current replica count for a marketplace tool
lift scale postgresql --path /opt/onelift/tools/postgresql

You can also scale marketplace tools directly from the OneLift dashboard under the tool's settings panel — no CLI needed.

Related

• lift scale - Scale replicas
• lift autoscale - Automatic scaling
• lift capacity - Server capacity
• lift limit - Resource limits
• lift update - Update marketplace tools to newer versions
• Zero-Downtime Deploys - Blue-green deployments