Skip to content

Automate Docker Deployments from CI/CD

git push → CI builds → BridgePort ships

The healthiest CI/CD split keeps builds where they belong — in CI, reproducible and versioned — and hands everything after the registry push to something that understands your servers. Doing that handoff with an SSH key in a CI secret and a bash script means your pipeline has root on production and no idea whether the deploy actually worked.

With BridgePort the handoff is a signed webhook. CI builds and pushes the image, then calls one endpoint; BridgePort pulls, deploys to the right servers, gates each step on health checks, rolls back automatically on failure, and reports the result — to your team and back to your tooling.

How it works

  1. CI builds and pushes

    GitHub Actions, GitLab CI, or any pipeline builds the image and pushes it to your registry — BridgePort works with the registries you already use.

  2. One webhook call

    The pipeline calls an incoming webhook, authenticated by HMAC signature (GitHub-style X-Hub-Signature-256 supported). Deploy a single service, or every service running that image.

  3. Health-gated rollout

    The deploy runs with the service’s strategy, or through a deployment plan that orders dependencies and checks health between steps.

  4. Rollback and report

    A failed health gate rolls the plan back automatically. Outgoing webhooks and notifications report success or failure to Slack, PagerDuty, or your own systems.

What you get

HMAC-signed webhooks

Incoming deploy triggers are authenticated by signature — no SSH keys or admin credentials in CI secrets.

Webhooks guide →

Deploy-all-for-image

One webhook can update every service running an image across the fleet — staging and production stay in step.

Incoming webhooks →

Plans with auto-rollback

Dependency-ordered, health-gated multi-service releases that undo themselves when a check fails.

Deployment plans →

Service accounts

Machine identities with scoped, role-capped API tokens — automation credentials that survive team turnover.

Service accounts →

Terraform provider

Manage servers, services, secrets, and config as code; runtime operations stay imperative and auditable.

Terraform provider →

Full API, SDK & CLI

Everything the UI does is a stable REST API — with a Go SDK, a CLI, and an MCP server for AI agents.

API reference →

Frequently asked questions

Does BridgePort build my images?

No — deliberately. CI owns the build so it stays reproducible and versioned; BridgePort owns everything after the registry push: pulling, deploying, health-checking, and rolling back.

How does CI authenticate to BridgePort?

Incoming webhooks are verified by HMAC signature (X-Webhook-Signature, or GitHub-compatible X-Hub-Signature-256). For richer automation, service accounts issue scoped API tokens with a role cap.

Are there ready-made CI examples?

Yes — the webhooks guide includes copy-paste examples for GitHub Actions, GitLab CI/CD, and a generic curl script.

Can I manage BridgePort itself as code?

Yes. The official Terraform provider manages configuration — servers, services, secrets, config files, registries — while deploys and restarts remain imperative operations through the UI, API, CLI, or webhooks.