Skip to main content
Focus: Understand deployment options for your Cube service, including always-on, on-demand, and scheduled modes, and how to configure them for your specific use case. Deployment options control when your Cube service runs and how it manages compute resources. Choose the right deployment mode based on your requirements for availability, cost, and use case.

​
Deployment modes overview

​
Available options

Metric Store supports three deployment modes:
  • Always on - Service runs continuously
  • On demand - Service starts when needed
  • Scheduled - Service runs on a time-based schedule
Each mode has different characteristics and use cases.

​
Always on deployment

​
Overview

Always on deployment keeps your Cube service running continuously, providing constant availability for queries and API access. Characteristics:
  • ✅ Service runs 24/7
  • ✅ Immediate availability for queries
  • ✅ No startup delays
  • ✅ Consistent performance
  • ⚠️ Higher compute costs
  • ⚠️ Resources always allocated

​
When to use

Production environments:
  • Critical business applications
  • High-traffic dashboards
  • Real-time analytics requirements
  • SLA-driven use cases
Always-available analytics:
  • Executive dashboards
  • Customer-facing analytics
  • Operational reporting
  • Time-sensitive queries

​
Configuration

  1. Access project settings
    • Click “Settings” button in header
    • Or click project name → “Edit project”
  2. Select deployment mode
    • Choose “Always on” from deployment options
    • Save changes
  3. Service starts
    • Service starts immediately
    • Heartbeat status reaches “RUNNING”
    • Available for queries right away
Production recommendationAlways on deployment is recommended for production environments where consistent availability and performance are critical.

​
On demand deployment

​
Overview

On demand deployment starts the Cube service when it’s needed and stops it when idle, optimizing resource usage and costs. Characteristics:
  • ✅ Cost-effective for low usage
  • ✅ Automatic start on first request
  • ✅ Automatic stop when idle
  • ⚠️ Startup delay on first request
  • ⚠️ Cold start time required

​
When to use

Development and testing:
  • Development environments
  • Testing and QA
  • Experimentation
  • Non-critical workloads
Low-traffic scenarios:
  • Occasional analytics needs
  • Ad-hoc reporting
  • Personal dashboards
  • Cost-sensitive use cases

​
Configuration

  1. Access project settings
    • Go to project settings
    • Navigate to deployment options
  2. Select deployment mode
    • Choose “On demand” from options
    • Save changes
  3. Service behavior
    • Service starts when first query arrives
    • Runs while active
    • Stops automatically after idle period
    • Restarts on next request
Startup timeOn demand deployment requires a startup time (typically 30-60 seconds) when the service starts. Users may experience a delay on the first request after an idle period.

​
Scheduled deployment

​
Overview

Scheduled deployment runs your Cube service on a specific schedule, such as business hours or specific days of the week. Characteristics:
  • ✅ Runs during specified times
  • ✅ Stops automatically outside schedule
  • ✅ Cost optimization for time-based needs
  • ✅ Predictable availability windows
  • ⚠️ Not available outside schedule
  • ⚠️ Requires schedule configuration

​
When to use

Business hours operations:
  • Office hours analytics
  • Business day reporting
  • Time-zone specific needs
  • Cost-optimized production
Scheduled reporting:
  • Daily reports
  • Weekly analytics
  • Monthly dashboards
  • Batch processing windows

​
Configuration

  1. Access project settings
    • Go to project settings
    • Select “Scheduled” deployment mode
  2. Configure schedule
    • Choose schedule type:
      • Weekdays - Monday-Friday with time range
      • Whole week - Monday-Sunday with time range
      • Custom - Cron-based scheduling
  3. Set time range
    • Start time - When service should start (e.g., 09:00)
    • End time - When service should stop (e.g., 18:00)
    • Timezone - Auto-detected or manually selected
  4. Save configuration
    • Schedule is saved and activated
    • Service starts/stops according to schedule

​
Schedule types

  • Weekdays
  • Whole week
  • Custom
Monday-Friday schedule:
  • Runs Monday through Friday
  • Configure start and end times
  • Example: 09:00 - 18:00 Monday-Friday
  • Stops automatically on weekends
Use cases:
  • Business hours operations
  • Office hours analytics
  • Workday reporting

​
Timezone configuration

Auto-detection:
  • System automatically detects browser timezone
  • Uses detected timezone for schedule
  • Convenient default for most users
Manual selection:
  • Override auto-detected timezone
  • Select from available timezones
  • Useful for remote teams or specific regions
Timezone format:
  • IANA timezone format (e.g., America/New_York)
  • Supports all major timezones
  • Handles daylight saving time automatically

​
Comparing deployment modes

​
Quick comparison

FeatureAlways OnOn DemandScheduled
Availability24/7When neededOn schedule
Startup timeNone30-60 secondsNone (within schedule)
CostHighestLowestMedium
Use caseProductionDevelopmentBusiness hours
PerformanceConsistentCold startsConsistent (within schedule)

​
Choosing the right mode

Select Always On if:
  • Production environment with high availability needs
  • Real-time analytics requirements
  • Consistent performance is critical
  • Cost is not a primary concern
Select On Demand if:
  • Development or testing environment
  • Low-traffic or occasional use
  • Cost optimization is important
  • Startup delays are acceptable
Select Scheduled if:
  • Business hours operations
  • Predictable usage patterns
  • Cost optimization for time-based needs
  • Availability during specific windows is sufficient

​
Best practices

​
Deployment strategy

Environment separation

Different modes per environmentUse Always On for production, On Demand for development, and Scheduled for staging.

Cost optimization

Right-size deploymentChoose deployment mode based on actual usage patterns, not just availability needs.

Monitor usage

Track patternsMonitor usage patterns to optimize deployment mode and schedule configuration.

Plan for growth

Scale appropriatelyStart with On Demand or Scheduled, upgrade to Always On as usage grows.

​
Configuration tips

  • Test deployment changes - Verify service starts/stops correctly
  • Monitor heartbeat status - Ensure service is running when expected
  • Set appropriate schedules - Align with actual usage patterns
  • Consider timezones - Configure schedules for your team’s timezone
  • Review periodically - Adjust deployment mode as needs change

​
Troubleshooting

​
Common issues

Possible causes:
  • Deployment mode configuration issues
  • Schedule not configured correctly
  • Environment or resource constraints
  • Service startup errors
Solutions:
  • Verify deployment mode is set correctly
  • Check schedule configuration (if scheduled)
  • Review environment configuration
  • Check server logs for errors
  • Try changing deployment mode
Possible causes:
  • Schedule configuration (if scheduled)
  • Idle timeout (if on demand)
  • Resource constraints
  • Service errors
Solutions:
  • Check schedule configuration and timezone
  • Verify on demand idle timeout settings
  • Review resource limits
  • Check service logs for errors
  • Consider switching to Always On
Possible causes:
  • Incorrect schedule configuration
  • Timezone mismatch
  • Cron expression errors
  • Service not respecting schedule
Solutions:
  • Verify schedule type and time range
  • Check timezone configuration
  • Validate cron expressions (if custom)
  • Test schedule configuration
  • Review schedule in project settings

Project Management

Manage projectsLearn how to configure deployment options in project settings.

Getting Started

Quick startUnderstand deployment options when creating your first project.