Skip to main content

Reliability & Performance

Built for production workloads. dbdeux is designed to be fast, reliable, and always available when your team needs it.

Availability​

Uptime SLA​

dbdeux targets 99.9% uptime for all production services:

  • Web application (IDE, DAG Explorer, dashboards)
  • API layer (run orchestration, notifications)
  • Compute engine (dbt execution)

Status Page​

Real-time status and incident history at dbdeux.datalakehouse.io/status:

  • Current status of all services
  • Historical uptime metrics
  • Incident timeline with root cause analysis
  • Scheduled maintenance notifications

Redundancy​

Every layer of the platform is designed for resilience:

  • Multi-zone deployment for high availability
  • Automated failover if any component becomes unhealthy
  • Zero-downtime deployments so updates never interrupt your work
  • Database replication with automatic failover

Performance​

IDE Responsiveness​

The Cloud IDE is optimized for speed:

  • Page load in under 2 seconds
  • Autocomplete suggestions in under 100ms
  • File switching is instantaneous (no reload)
  • Large files (10,000+ lines) render smoothly

Run Execution​

Runs start fast and execute efficiently:

  • Cold start: Under 10 seconds from click to execution
  • Warm start: Under 3 seconds for back-to-back runs
  • Parallel execution: Multiple models run concurrently (thread count configurable)
  • Streaming logs: See output in real-time, not after completion

DAG Rendering​

The lineage graph handles projects of any size:

  • 100 models: Instant render
  • 1,000 models: Renders in under 1 second
  • 5,000+ models: Progressive loading with instant interaction

Scalability​

Horizontal Scaling​

dbdeux scales automatically based on demand:

  • Compute containers scale up during peak hours and down during quiet periods
  • No manual capacity planning required
  • No performance degradation during traffic spikes
  • Cost-efficient (you only consume resources when running jobs)

Concurrent Runs​

Run as many jobs as you need simultaneously:

  • Multiple team members running models at the same time
  • Scheduled jobs executing in parallel
  • No queueing or waiting (each run gets its own isolated container)

Large Projects​

Tested and optimized for enterprise-scale dbt projects:

  • Thousands of models in a single project
  • Complex dependency graphs with deep nesting
  • Large SQL files with extensive Jinja logic
  • Hundreds of sources and seeds

Disaster Recovery​

Data Durability​

All platform data (metadata, configurations, run history) is:

  • Replicated across multiple availability zones
  • Backed up continuously with point-in-time recovery
  • Retained according to your plan's data retention policy

Recovery Time​

In the unlikely event of a major incident:

  • RTO (Recovery Time Objective): 1 hour
  • RPO (Recovery Point Objective): 5 minutes

Your Data is Safe Regardless​

Because dbdeux operates on a metadata-only model, a platform outage never puts your warehouse data at risk. Your data lives exclusively in your warehouse, which has its own independent backup and recovery mechanisms.