For Agents
Promote releases through staged delivery pipelines on GKE, Cloud Run, and Anthos, with approval gates, rollback, and rollout management. Useful for managed continuous-delivery workflows on GCP.
Get started with Cloud Deploy API in minutes using your preferred integration method.
# Add to your MCP client config (Claude Desktop, Cursor, Windsurf)
{
"jentic": {
"url": "https://api.jentic.com/mcp",
"auth": "oauth"
}
}
# Then ask your agent:
"promote a google cloud deploy release"
# → Jentic returns the GET /events tool with parameter schema, agent executes.What an agent can do with Cloud Deploy API API.
Create a delivery pipeline that defines the sequence of target environments
Cut a release that promotes a specific container image through the pipeline
Approve, advance, or abandon a pending rollout
Roll back to a previous successful release on a given target
GET STARTED
Use for: I need to promote the latest release from staging to production, Approve the pending rollout waiting on the production target, Roll back the production deployment to the previous successful release, List every release on our service in the last 30 days
Not supported: Does not build artefacts, run application traffic, or store source code — use for staged release promotion, approval, and rollback only.
Google Cloud Deploy is a managed, opinionated continuous-delivery service that promotes container images and Kubernetes manifests through a series of target environments such as staging and production. The API exposes delivery pipelines, releases, rollouts, targets, and approval steps so agents can promote a release to the next stage, abandon a release, approve a pending rollout, or roll back to a previous successful deployment. Cloud Deploy integrates with Cloud Build for build artefacts and with GKE, Cloud Run, and Anthos for deployment targets.
Cancel a rollout that is mid-flight on a target environment
List targets, releases, and rollouts to surface delivery state to dashboards
Configure custom target types for delivery to non-default runtimes
Patterns agents use Cloud Deploy API API for, with concrete tasks.
★ Staged Release Promotion to GKE
Platform teams use Cloud Deploy to promote a tested container image through dev, staging, and production GKE clusters with explicit approval gates between stages. Each promotion is an advanceRollout call; production rollouts wait on an approveRollout call from a release manager. Setup typically takes a day per service; subsequent releases take seconds.
Call POST /v1/{name}:advance on the rollout resource with phaseId set to the target phase, then POST /v1/{name}:approve once production is ready
Automated Rollback on Failure
SRE workflows watch for elevated error rates after a rollout and call rollbackTarget to restore the previous release. The API returns a new rollout that re-applies the prior image and Kubernetes manifests, so service health returns to baseline without a manual rerun of the original pipeline.
Call POST /v1/{name}:rollbackTarget with the target resource name and an optional release reference to specify which release to roll back to
Release Awaiting-Approval Dashboard
Release managers use the rollouts list endpoint with a filter on approvalState to surface every rollout currently waiting for sign-off. The dashboard pairs the data with deployment metadata so reviewers can approve or abandon with full context. The same endpoint backs alerting that escalates if a rollout stalls past a threshold.
Call GET on the rollouts collection with filter=approvalState=NEEDS_APPROVAL across the configured delivery pipelines
AI Agent Release Approver via Jentic
An agent invoked by an on-call engineer searches Jentic for the right Cloud Deploy operation, gathers the rollout context, and either advances, approves, or rolls back the release based on the engineer's intent. Jentic injects a scoped OAuth token at execution time so the agent never holds long-lived credentials.
Search Jentic for 'approve google cloud deploy rollout', execute against POST /v1/{name}:approve with approved=true, and report the resulting rollout state
32 endpoints — google cloud deploy is a managed, opinionated continuous-delivery service that promotes container images and kubernetes manifests through a series of target environments such as staging and production.
METHOD
PATH
DESCRIPTION
/v1/{+name}:advance
Advance a rollout to its next phase
/v1/{+name}:approve
Approve a rollout awaiting sign-off
/v1/{+name}:abandon
Abandon a release that should no longer be promoted
/v1/{+name}:cancel
Cancel an in-flight rollout
/v1/{+name}:rollbackTarget
Roll back a target to a previous release
/v1/{+name}/operations
List long-running operations on a resource
/v1/{+name}/locations
List Cloud Deploy locations available to a project
/v1/{+name}:advance
Advance a rollout to its next phase
/v1/{+name}:approve
Approve a rollout awaiting sign-off
/v1/{+name}:abandon
Abandon a release that should no longer be promoted
/v1/{+name}:cancel
Cancel an in-flight rollout
/v1/{+name}:rollbackTarget
Roll back a target to a previous release
Three things that make agents converge on Jentic-routed access.
Credential isolation
Google Cloud OAuth 2.0 credentials and service account keys are stored encrypted in the Jentic vault. Agents receive short-lived access tokens scoped to cloud-platform — raw service account JSON never enters the agent context.
Intent-based discovery
Agents search Jentic by intent (e.g. 'approve google cloud deploy rollout' or 'rollback cloud deploy target') and Jentic returns the matching operation with its parameter schema, so the agent can call the right endpoint without reading Discovery docs.
Time to first call
Direct Cloud Deploy integration: 2-5 days for OAuth, pipeline modelling, and long-running operation polling. Through Jentic: under an hour — search, load schema, execute.
Alternatives and complements available in the Jentic catalogue.
Cloud Build API
Produces the container artefacts that Cloud Deploy promotes through environments
Use Cloud Build to build and test; use Cloud Deploy to promote the resulting image through staging and production
Kubernetes Engine API
Manages the GKE clusters that serve as Cloud Deploy targets
Use GKE API to provision or scale the cluster; use Cloud Deploy to promote workloads onto it
Cloud Run Admin API
Manages Cloud Run services that Cloud Deploy can target as a runtime
Use Cloud Run API for direct service config; use Cloud Deploy when you need staged promotion across Cloud Run environments
Specific to using Cloud Deploy API API through Jentic.
What authentication does the Cloud Deploy API use?
The Cloud Deploy API uses OAuth 2.0 with the cloud-platform scope. Tokens are issued for a Google service account or end user. Through Jentic, the OAuth credential lives in the Jentic vault and the agent receives a short-lived access token only — the underlying service account JSON never enters agent context.
Can I promote a release to the next environment with this API?
Yes. Call POST /v1/{name}:advance on the rollout resource to move it to the next phase, or create a new rollout against the next target if the pipeline is configured for serial promotion. Production rollouts that require approval pause on a NEEDS_APPROVAL state until /v1/{name}:approve is called.
What are the rate limits for the Cloud Deploy API?
Cloud Deploy quotas are published per-method in the Google Cloud console under the Cloud Deploy quota page; typical defaults are several hundred read requests per minute per project, with concurrent rollout limits per target type (GKE, Cloud Run, custom).
How do I roll back a production release through Jentic?
Search Jentic for 'rollback google cloud deploy target' and execute against POST /v1/{name}:rollbackTarget on the target resource. The API creates a new rollout that re-applies the previous successful release; the agent can poll the rollout for completion and report the resulting state.
Does Cloud Deploy support custom deployment targets?
Yes. Custom target types are first-class resources — define a customTargetType with a render and deploy action, and reference it from a target inside a delivery pipeline. The same approval, advance, abandon, and rollback verbs apply to custom targets.
Is Cloud Deploy free?
Cloud Deploy charges a per-target-environment monthly fee documented on the Cloud Deploy pricing page. Builds run by Cloud Build and the runtime resources targeted (GKE, Cloud Run) are billed separately under their own services.
/v1/{+name}/operations
List long-running operations on a resource
/v1/{+name}/locations
List Cloud Deploy locations available to a project