Platform capabilities
Every module, every option — from the first deployment to single sign-on. All of it included, for every team.
01
Deployments
From your repository or registry to a running, healthy container — automatically.
- Sources: GitHub (GitHub App), Azure DevOps, GitLab, Gitea / Forgejo, Bitbucket, any Git URL, Docker images and Docker Compose
- Build packs: Dockerfile, Nixpacks, Node.js, Bun, Python, Go, static sites and shell scripts — detected automatically
- Push-to-deploy with webhooks, pull request previews with their own domain, deploy hooks for any CI
- Deployment status on commits and pull requests (GitHub, Azure DevOps)
- Zero-downtime rollouts: traffic moves only after the health check passes
- The port comes from the image (EXPOSE): set automatically until the first successful deployment, then any mismatch is reported
- Canary deployments: 5–50 % of the traffic goes to the new version, promoted or rolled back automatically
- Scheduled jobs (cron) per application, with run history, output, timeouts and failure alerts
- Git over SSH for GitLab, Bitbucket and Gitea on premises, with a deploy key per application
- Pre- and post-deploy commands, replicas, CPU and memory limits
- Watch paths for monorepos, several servers per application, dedicated build servers
- Private registries: Docker Hub, GHCR, GitLab, Amazon ECR, Harbor, Quay and any OCI registry
02
Versions and rollbacks
Every release kept, every rollback instant.
- Every production deployment becomes a numbered version (v1, v2…)
- The live version and the last 5 are kept for 15 days, older ones removed automatically
- Roll back by drag and drop or in two clicks — no rebuild, current variables
- The old version stops only once the new one is healthy; approvals apply
- Docker Compose applications keep the image of every service and the compose file
03
Domains and HTTPS
Traefik with automatic certificates, configured for you.
- Automatic Let's Encrypt certificates, renewal and an expiry watchdog every 6 hours
- Wildcard certificates with DNS-01: Cloudflare, Route 53, Azure DNS, Google Cloud DNS, DigitalOcean, Hetzner, OVH, GoDaddy, RFC 2136…
- Uploaded certificates (corporate CA, EV / OV) distributed to every proxy
- HTTP → HTTPS and www redirects, a port and a path per domain, search indexing control
- DNS checks against the server's IP, with Cloudflare proxy detection
- Access protection per application: single sign-on with PortDock, password, allowed addresses and a rate limit
- Changes applied in seconds, without a redeploy
04
Databases and services
Managed databases with real backups, and ready-made services.
- PostgreSQL, MySQL, MariaDB, Microsoft SQL Server, MongoDB, Redis, Valkey and ClickHouse
- Online backups following each engine's documentation: manual, scheduled or all at once
- Restore in one click, after an automatic safety backup; test restores that prove a backup works
- Point-in-time recovery for PostgreSQL, to any second of the recovery window
- PostgreSQL replicas on the same or another server, with manual or automatic failover
- Retention and upload to S3 (AWS, Cloudflare R2, Backblaze B2, MinIO)
- Access from outside by domain or public IP
- SQL console in the browser for PostgreSQL, MySQL, MariaDB, SQL Server and ClickHouse — read-only for members
- 31 one-click services (Keycloak, Nextcloud, n8n, Gitea, Grafana, Jenkins, Open WebUI…) and a catalog of 127 Docker images
- Service catalogs: the community catalog, Coolify's 350+ templates or your own
- Import from Coolify, Heroku and Dokku: applications, databases and services
05
Configuration
Variables and settings that are safe and easy to change.
- Environment variables encrypted with AES-256-GCM, for the build and the runtime
- Variables per Docker Compose service, with the values of docker-compose.yml shown and overridable
- Apply new variables with a restart, without a rebuild
- Separate variables for pull request previews
- Shared variables for the organization, a project or an environment, with a comparison between environments
- Secrets read from HashiCorp Vault or Azure Key Vault when a deployment starts, never stored in PortDock
- Persistent storage: Docker volumes kept across deployments, read-only mounts, server directories
06
Servers and containers
Your whole fleet, from one dashboard.
- The local server and remote servers over SSH, with encrypted keys and host key pinning
- One-click setup: Docker, the network and the Traefik proxy
- Containers, images and volumes of every server in one view, with CPU, memory and the resource they belong to
- Web terminal, live logs, start / stop / restart, images, volumes, networks and prune
- Live topology map with crash-loop detection, in the style of Argo CD
07
Kubernetes
Your cluster as a server — or a new one, set up in one command.
- A namespace of any cluster as a server, connected with a service account limited to it
- An application becomes a Deployment with a rolling update gated by startup and readiness probes, a Service and an Ingress
- A cert-manager certificate per HTTPS domain, a Secret of variables per deployment, a PersistentVolumeClaim per volume
- Pre-deploy commands and scheduled jobs run as Kubernetes Jobs; versions and rollbacks without a rebuild
- Images built on a Docker server and pushed to the cluster's registry, pulled with a secret PortDock creates
- Logs, terminal, start / stop / restart, containers, topology and metrics as on Docker servers
- All in one on a fresh server: k3s, cert-manager with Let's Encrypt, a private registry, Docker for the builds and PortDock with test and prod — one command
08
Monitoring and alerts
Know before your users do.
- CPU, memory, disk, network and load history for servers and every resource
- Alert rules with a duration and a severity
- Notifications by email, Slack, Microsoft Teams, Discord, Telegram or signed webhook
- Alerts for deployments, backups, certificates, servers and updates
- Prometheus endpoint, and log drains to Grafana Loki, Elasticsearch, Graylog, Splunk, syslog or HTTP
- OpenTelemetry settings injected into applications, for traces, metrics and logs
09
Uptime and status page
Monitors for everything that must stay up.
- Monitors for sites and APIs: status code, keyword, response time and TLS certificate expiry
- Incidents opened and resolved automatically, with notifications
- A public status page at /status with 90 days of uptime
10
Security scanning
Three scanners, one policy, reports for every deploy.
- Vulnerability scanning of images with Trivy, optionally blocking on critical or high findings
- Static analysis of the source with Semgrep (SAST), right after the checkout
- Dynamic scan of the running application with OWASP ZAP (DAST)
- Reports for every scan: HTML, CSV, JSON, SARIF and the ZAP report, one by one or as a ZIP
- A ZIP with every report of a deployment, from its page
- Releases signed with Sigstore, with an SBOM and SLSA provenance; updates verify the signature
11
Teams and access
The right people, with the right rights.
- Owner, admin and member roles, plus viewer, developer and maintainer per project
- Protected environments and deployment approvals (four-eyes principle)
- Two-factor authentication with recovery codes and an organization-wide policy
- Single sign-on (OIDC) and LDAP / Active Directory
- Audit log with CSV / JSON export and SIEM streaming (webhook or syslog)
- API tokens with a scope (read, deploy, write, full), limited to projects, with an expiry date; service accounts for CI
- Compliance report for auditors: access, approvals, scans, backups and incidents of a period
12
Corporate networks
Made for networks behind a proxy, or without internet.
- HTTP(S) proxy with user and password for PortDock, builds and containers
- NO_PROXY with names, domains, IP addresses and CIDR ranges
- Docker Hub mirror (Nexus, Artifactory, Harbor) applied to each server's daemon without stopping containers
- npm, PyPI, Go and Maven mirrors for builds, and a mirror of the Trivy database
- A per-application choice: follow the setting, always use the proxy or never
13
Configuration as code and API
Everything the interface does, from a file, a terminal or an AI agent.
- portdock.yaml declares applications, domains, variables, storage, jobs and access protection; pdctl apply shows the plan and changes only what differs
- pdctl, the command-line client, for Linux, macOS and Windows: deploy --wait, env, jobs, run, export
- A Terraform provider for projects, environments, applications, domains and variables
- OpenAPI 3.1 description of every endpoint, browsable in the interface
- MCP server: Claude Code, VS Code, Cursor and other agents deploy, roll back and read logs with an API token
14
Platform
The details that make it pleasant every day.
- One-click updates from GitHub releases, with a backup, a signature check and automatic rollback, on Docker and in Kubernetes
- Instance backup: PortDock's own database and key, encrypted, on a schedule; restore on a new server from the first-run page
- AI assistant that reads the platform with your permissions and proposes actions you confirm, with any model on OpenRouter or a local server
- Interface in English, Romanian and Russian, with a light and a dark theme
- Command palette (Ctrl K), real-time interface and a full-width responsive layout
- REST API with tokens and a single binary with the interface embedded
Evaluate PortDock on your own infrastructure
Installation takes a few minutes on any Linux server. No licence key or registration is required.
$ git clone https://github.com/gavrilencu/PortDock.git portdock$ cd portdock$ sudo bash install.sh