Skip to main content

Frequently Asked Questions

Find answers to the most common questions about Harness Chaos Engineering, from getting started to advanced implementation topics.

Getting Started​

What is Chaos Engineering?​

Chaos Engineering is the discipline of experimenting on a system to build confidence in the system's capability to withstand turbulent conditions in production. It involves intentionally introducing failures in a controlled manner to identify weaknesses before they cause outages.

Why should I use Chaos Engineering?​

Chaos Engineering helps you:

  • Identify weaknesses before they cause production outages
  • Build confidence in your system's resilience
  • Improve incident response through practice and preparation
  • Reduce MTTR (Mean Time To Recovery) during real incidents
  • Validate disaster recovery procedures and assumptions

Is Chaos Engineering safe?​

Yes, when done properly. Chaos Engineering emphasizes:

  • Controlled experiments with defined blast radius
  • Comprehensive monitoring and automatic rollback
  • Starting small and gradually increasing scope
  • Safety measures and circuit breakers
  • Team coordination and communication

How is Harness Chaos Engineering different from other tools?​

Harness Chaos Engineering offers:

  • Enterprise-grade platform with advanced security and compliance
  • Comprehensive fault library covering infrastructure, applications, and cloud services
  • Native CI/CD integration for continuous resilience testing
  • Advanced automation with intelligent rollback and remediation
  • Team collaboration features including GameDays and shared experiments

Implementation​

Where should I start with Chaos Engineering?​

Follow this progression:

  1. Start in staging environments, never production initially
  2. Begin with simple experiments like pod restarts or CPU stress
  3. Use small blast radius affecting only a few instances
  4. Set up comprehensive monitoring before running experiments
  5. Gradually expand scope and complexity as confidence grows

What experiments should I run first?​

Recommended first experiments:

  • Pod/Container deletion - Test application restart resilience
  • CPU stress - Validate auto-scaling and performance under load
  • Network latency - Test timeout handling and user experience
  • Database connection failures - Validate connection pooling and failover
  • Memory pressure - Test memory management and garbage collection

How do I choose what to test?​

Consider these factors:

  • Business criticality - Start with most important user journeys
  • Historical incidents - Test scenarios based on past outages
  • System dependencies - Focus on single points of failure
  • Recent changes - Validate resilience of new features or infrastructure
  • Compliance requirements - Test disaster recovery procedures

How often should I run chaos experiments?​

Frequency depends on your maturity:

  • Getting started: Weekly experiments in staging
  • Intermediate: Daily automated experiments, weekly production tests
  • Advanced: Continuous chaos testing integrated into CI/CD
  • Mature: Real-time chaos testing with intelligent automation

Technical Questions​

What platforms does Harness Chaos Engineering support?​

We support:

  • Kubernetes (all major distributions: EKS, AKS, GKE, OpenShift)
  • Linux (Ubuntu, RHEL, CentOS, Debian, Fedora)
  • Cloud platforms (AWS, Azure, GCP)
  • Windows (limited fault types)
  • VMware vSphere
  • Container platforms (Docker, containerd, CRI-O)

Can I run chaos experiments in production?​

Yes, but with proper precautions:

  • Start with staging and build confidence first
  • Use small blast radius (1-5% of instances initially)
  • Run during low-traffic periods to minimize user impact
  • Have comprehensive monitoring and automatic rollback
  • Coordinate with teams and stakeholders
  • Follow change management procedures

How do I measure the success of chaos experiments?​

Key metrics include:

  • Resilience score - Overall system resilience rating
  • Recovery time - How quickly systems return to normal
  • Error rates - Impact on user-facing errors
  • Performance metrics - Response time and throughput impact
  • Business metrics - Revenue, user experience, SLA compliance

What happens if an experiment goes wrong?​

Harness Chaos Engineering includes multiple safety mechanisms:

  • Automatic rollback when thresholds are breached
  • Manual stop capability at any time
  • Blast radius limits to contain impact
  • Comprehensive monitoring to detect issues quickly
  • Team notifications for immediate response

Security and Compliance​

Is Harness Chaos Engineering secure?​

Yes, we implement enterprise-grade security:

  • SOC 2 Type II certified
  • ISO 27001 compliant
  • End-to-end encryption for data in transit and at rest
  • RBAC (Role-Based Access Control) for fine-grained permissions
  • Audit logging for all activities and changes
  • SAML/SSO integration with enterprise identity providers

Can I use Chaos Engineering in regulated industries?​

Absolutely. Many regulated industries use chaos engineering:

  • Financial services - Test trading systems and payment processing
  • Healthcare - Validate patient data systems and medical devices
  • Government - Test critical infrastructure and citizen services
  • Aerospace - Validate flight systems and ground operations

Key considerations:

  • Compliance documentation - Maintain detailed experiment records
  • Change management - Follow established change control procedures
  • Risk assessment - Evaluate and document potential impacts
  • Stakeholder approval - Get necessary approvals before testing

How do you handle data privacy?​

We protect data privacy through:

  • Data minimization - Collect only necessary experiment data
  • Encryption - All data encrypted in transit and at rest
  • Access controls - Strict access controls and audit trails
  • Data retention - Configurable retention policies
  • Regional compliance - Support for GDPR, CCPA, and other regulations

Troubleshooting​

My experiment won't start. What should I check?​

Common issues and solutions:

  1. Permissions - Verify RBAC permissions for target resources
  2. Connectivity - Check network connectivity to target systems
  3. Resource availability - Ensure target resources exist and are healthy
  4. Configuration - Validate experiment configuration and parameters
  5. Dependencies - Check if required services or tools are running

Experiments are failing unexpectedly. How do I debug?​

Debugging steps:

  1. Check experiment logs for error messages and details
  2. Verify target system health before and during experiments
  3. Review monitoring data for system behavior patterns
  4. Validate success criteria and thresholds
  5. Test in isolation to identify specific failure points

How do I handle false positives in monitoring?​

Strategies for reducing false positives:

  • Tune thresholds based on historical data and normal variation
  • Use multiple metrics to validate actual impact
  • Implement grace periods to account for temporary fluctuations
  • Regular threshold review and adjustment based on system changes
  • Composite health checks combining multiple indicators

My team is resistant to Chaos Engineering. How do I get buy-in?​

Building team support:

  • Start with education about chaos engineering benefits
  • Share success stories from other organizations
  • Begin with low-risk experiments to build confidence
  • Involve skeptics in experiment design and execution
  • Demonstrate value through improved incident response
  • Address concerns about safety and potential impact

Advanced Topics​

Can I create custom faults?​

Yes, through several methods:

  • BYOC (Bring Your Own Chaos) - Upload custom fault scripts

  • API integration - Use REST APIs to trigger custom failures

  • Custom probes - Create application-specific health checks

  • Script execution - Run custom scripts during experiments

How do I integrate with my existing monitoring tools?​

Harness integrates with popular tools:

  • Prometheus/Grafana - Native metrics integration
  • Datadog - Real-time monitoring and alerting
  • New Relic - APM and infrastructure monitoring
  • Splunk - Log analysis and correlation
  • PagerDuty - Incident management and escalation

Can I automate chaos experiments in CI/CD pipelines?​

Yes, we provide:

  • GitHub Actions integration
  • Jenkins plugins and pipeline steps
  • GitLab CI integration
  • Azure DevOps extensions
  • Native Harness CI/CD integration
  • REST API for custom integrations

How do I scale chaos engineering across multiple teams?​

Scaling strategies:

  • Standardized templates for common experiment types
  • Shared experiment library for reusable scenarios
  • Team-specific permissions and resource isolation
  • Centralized reporting and analytics
  • Training programs and best practice sharing
  • Center of Excellence for chaos engineering governance

Pricing and Licensing​

How is Harness Chaos Engineering priced?​

Pricing is based on:

  • Infrastructure targets - Number of nodes/instances under test
  • Experiment executions - Number of experiments run per month
  • Advanced features - Enterprise features and support levels
  • Deployment model - SaaS vs. Self-Managed pricing

Contact sales for detailed pricing information.

Is there a free tier available?​

Yes, we offer:

  • Free tier with limited experiments and targets
  • Trial period for full feature evaluation
  • Developer accounts for learning and experimentation
  • Open source components for basic chaos testing

What support options are available?​

Support tiers include:

  • Community support - Forums and documentation
  • Standard support - Business hours email support
  • Premium support - 24/7 support with SLA guarantees
  • Enterprise support - Dedicated customer success manager

Getting Help​

Where can I find more information?​

Resources available:

How do I contact support?​

Support channels:

  • In-app support - Submit tickets directly from the platform
  • Email support - support@harness.io
  • Phone support - Available for premium customers
  • Emergency support - 24/7 for critical issues (enterprise customers)

Can I request new features?​

Yes, we welcome feedback:

  • Feature requests through in-app feedback
  • Community voting on proposed features
  • Direct feedback to customer success managers
  • Beta programs for early access to new features

Still have questions?​

If you can't find the answer you're looking for:

  1. Search our Knowledge Base for detailed articles
  2. Join our Community Forum to ask questions
  3. Contact Support for technical assistance
  4. Schedule a demo to speak with our experts
Quick Help

For immediate assistance, use the in-app help chat or check our Resources section for common issues and solutions.