Dyrector

Dyrector is a self-hosted continuous delivery and deployment platform designed to orchestrate containerized workloads across infrastructure nodes. It replaces direct SSH access and ad-hoc scripts with a central management plane and node agents communicating over mutual TLS via gRPC. Teams host the control plane and database entirely on their own compute, keeping deployment credentials, version configurations, and container environments under strict internal control.

Where Dyrector fits

Dyrector operates in the release orchestration layer between container registries and target infrastructure. Rather than functioning as a full platform-as-a-service that compiles source code or routes traffic, it focuses on version management, configuration updates, and workload rollouts. In privacy-conscious organizations, routing deployment tokens and cluster topologies through hosted delivery services introduces unnecessary operational exposure. Dyrector eliminates this by running the dashboard, API, and PostgreSQL database on internal infrastructure, communicating with worker nodes through outbound agent connections.

This model sits between basic container managers like Portainer and declarative GitOps tools like ArgoCD. Where container managers focus on single-host debugging and GitOps engines demand that every modification pass through Git commits, Dyrector provides an interactive deployment console. Platform teams register Docker hosts, Podman machines, or Kubernetes clusters as deployment targets, then allow authorized developers to update container configurations, promote versions, and inspect runtime state without distributing shell or cluster credentials.

What it does well

• Node agent separation: Dedicated Go agents—dagent for Docker or Podman and crane for Kubernetes—initiate outbound gRPC connections to the platform, avoiding inbound firewall openings.
• Multi-environment version tracking: Maintains a structured ledger of container versions and configuration states across test, staging, and production environments from a unified console.
• Self-service test environments: Allows developers and QA engineers to stand up branch-specific instances on available compute without filing infrastructure tickets.
• Centralized configuration and secret handling: Employs a structured JSON editor and reusable bundle configurations to inject environment variables and secrets into containers across multiple targets.