CI/CD for Robots
Roboticks CI/CD brings software delivery best practices to robotics. Package applications as versioned capsules, run hardware-in-the-loop tests on managed cloud runners or self-hosted infrastructure, then deploy with progressive rollout strategies that include automatic health checks and rollback.
Pipeline Flow
The Roboticks deployment pipeline follows a proven flow: Code your robot application, package it as a Capsule with versioning and integrity checks, Test on cloud or local runners with hardware-in-the-loop execution, Rollout using canary, blue-green, progressive, or all-at-once strategies, then Monitor with real-time observability and AI diagnostics.
Capsule Archives
A capsule is the versioned deployment unit — a packaged robot application supporting .tar.gz, .zip, .tar, and .tgz formats. Each capsule uses semantic versioning (major.minor.patch), S3 storage with SHA256 integrity checksums, dependency management between capsules, and CI/CD source tracking for repository, branch, and commit hash.
Rollout Strategies
Five deployment strategies cover every scenario: All At Once for simultaneous deployment to all devices; Progressive for staged percentages with configurable wait times between stages; Canary for validating on a small test group before full fleet rollout; Blue-Green for 50/50 A/B testing with instant rollback; and Offline for generating download links for air-gapped devices.
Cloud Runners
Managed test infrastructure runs on AWS EC2 with multiple architecture categories: ARM, x86, Accelerated ARM (GPU), and Accelerated x86 (GPU). Instance tiers range from Micro to XLarge, with NVIDIA T4G and A10G GPUs available for accelerated workloads. Billing is per compute unit, varying by instance type.
Local Runners
Self-hosted runners operate as long-lived polling agents that pull test jobs from the platform. They use Docker-based execution environments with token-based authentication and rotation support. Configure concurrent job capacity to match your infrastructure. Registration tokens simplify fleet management of runner agents.
Traditional Deployment vs Roboticks CI/CD
| Aspect | DIY / Traditional | Roboticks |
|---|---|---|
| Packaging | Manual builds, SCP to devices | Versioned capsules with SHA256 integrity and dependency tracking |
| Testing | Test on physical devices in the lab | Cloud runners (ARM/x86/GPU) + self-hosted runners with HIL support |
| Rollout | All-or-nothing deployment scripts | Canary, blue-green, progressive, and offline strategies |
| Rollback | Manual revert with downtime | Automatic rollback on health check failure |
| Traceability | Spreadsheets or wiki pages | Full CI/CD source tracking — repo, branch, commit, pipeline URL |
Documentation
For setup guides, API references, and technical details, visit the full documentation.