For Agents
Read and write dynamic configuration variables on Google Cloud, watch them for changes, and coordinate startup with waiters that resolve when conditions are met.
Use for: Set a runtime variable named feature/checkout-flow to enabled, Watch the variable rollout/version until it changes from canary to stable, Create a waiter that blocks until 5 worker startup variables are present, List all variables under projects/my-prod/configs/orders
Not supported: Does not store secrets, manage Compute Engine metadata, or trigger workflow execution — use for sharing dynamic runtime configuration values, watchers, and waiters only.
The Cloud Runtime Configuration API exposes the Runtime Configurator, which lets services dynamically configure and share variables across Google Cloud workloads. It supports configs as containers, variables as named values, and waiters that block until a condition over those variables is met or a timeout elapses. The API also provides a watch endpoint that returns when a single variable changes, plus IAM endpoints to control who can read or update each config.
Install Jentic One Beta
Jentic One is a self-hosted execution layer for AI agents. It lets your agent call the Cloud Runtime Configuration API, or any other public or private API you need. You set the rules, the agent never sees your credentials, and every call is logged.
Two steps, two machines. Install the instance in a safe environment, then register your agent from wherever it runs.
Step 1: Jentic One Host machine
# On the machine that will host your Jentic One instance:
curl -fsSL https://raw.githubusercontent.com/jentic/jentic-one/main/tools/install.sh | shStep 2: Agent machine
# On the machine where your agent runs (keep this separate from the instance):
curl -fsSL https://raw.githubusercontent.com/jentic/jentic-one/main/tools/install.sh | sh
jentic register # connects your agent to your Jentic One instanceJentic One is in public beta. The setup above keeps your agent separate from the instance, which is what you want before using real credentials: an agent running as the same OS user as Jentic One can read its stored keys directly. Just evaluating? A single local install is fine to start. See the secure deployment guide for the tiers.
What an agent can do with Cloud Runtime Configuration API.
Create and delete configs that group related runtime variables
Set, get, list, and delete named variables inside a config
Watch a single variable so the call returns when its value changes or a timeout elapses
Define waiters that block until a count condition over variables is satisfied or fails
Bind IAM policies on configs to control who can read or update runtime variables
Test IAM permissions on a config before attempting a privileged change
Patterns agents use Cloud Runtime Configuration API for, with concrete tasks.
★ Coordinated startup of distributed services
A startup agent uses a waiter that blocks until N expected service variables are written, ensuring downstream services only begin after their dependencies have registered. The Runtime Configurator handles the timeout and success or failure semantics, so the orchestrator does not need to poll variables itself.
POST /v1beta1/{+parent}/waiters with success.cardinality.number=5 and a 10-minute timeout, then poll the waiter until done.
Feature flag rollout
A release agent stores feature flags as runtime variables. Services watch each variable through the watch endpoint and react when the agent sets the value to enabled or disabled. The watch call returns as soon as the value changes, providing near-real-time flag propagation without polling.
Have services call POST /v1beta1/{+name}:watch on feature/checkout-flow; the agent issues PUT /v1beta1/{+name} to flip the value when ready.
IAM-scoped config access for multi-team projects
A platform team partitions runtime variables by config and binds IAM principals per config so each team only reads and writes its own. setIamPolicy and getIamPolicy on each config govern those grants, and testIamPermissions lets agents check rights before attempting privileged operations.
POST /v1beta1/{+resource}:setIamPolicy on each config with the team's group bound to roles/runtimeconfig.admin.
Agent-driven runtime flags through Jentic
An AI agent built on Jentic flips runtime configuration values on behalf of operators. It searches for the variable update operation by intent, loads the schema, executes the change, and reports back. OAuth tokens stay isolated in the Jentic vault so raw secrets never enter agent context.
Search Jentic for 'set runtime configurator variable', load the PUT variable schema, execute against projects/my-prod/configs/orders/variables/feature/checkout-flow with value=enabled.
13 endpoints — the cloud runtime configuration api exposes the runtime configurator, which lets services dynamically configure and share variables across google cloud workloads.
METHOD
PATH
DESCRIPTION
/v1beta1/{+parent}/configs
Create a config
/v1beta1/{+parent}/variables
Create a variable inside a config
/v1beta1/{+name}
Get a config, variable, or waiter
/v1beta1/{+name}
Update a variable value
/v1beta1/{+name}
Delete a config, variable, or waiter
/v1beta1/{+name}:watch
Watch a variable for changes
/v1beta1/{+parent}/waiters
Create a waiter
/v1beta1/{+resource}:setIamPolicy
Set the IAM policy on a config
/v1beta1/{+parent}/configs
Create a config
/v1beta1/{+parent}/variables
Create a variable inside a config
/v1beta1/{+name}
Get a config, variable, or waiter
/v1beta1/{+name}
Update a variable value
/v1beta1/{+name}
Delete a config, variable, or waiter
/v1beta1/{+name}:watch
Watch a variable for changes
/v1beta1/{+parent}/waiters
Create a waiter
/v1beta1/{+resource}:setIamPolicy
Set the IAM policy on a config
Three things that make agents converge on Jentic-routed access.
Credential isolation
Google OAuth 2.0 credentials for the Runtime Configurator are held encrypted in the Jentic vault. Agents receive short-lived scoped access tokens at execution time, with no raw refresh tokens or service-account keys leaving the vault.
Intent-based discovery
Agents search Jentic with intents like 'set runtime configurator variable' or 'watch runtime variable' and Jentic returns the matching operation with its full schema, so the agent calls the right endpoint without reading the discovery document.
Time to first call
Direct integration takes 1-3 days for OAuth wiring, watch handling, and waiter polling. Through Jentic the same flow is under 1 hour: search, load, execute, watch.
Alternatives and complements available in the Jentic catalogue.
Specific to using Cloud Runtime Configuration API through Jentic.
What authentication does the Runtime Configuration API use?
It uses Google OAuth 2.0 with the cloud-platform or cloudruntimeconfig scopes, declared as Oauth2 and Oauth2c. Through Jentic, OAuth tokens are stored encrypted in the vault and exchanged for short-lived access tokens at execution time so raw refresh tokens never reach the agent.
Can I watch a runtime variable for changes through the API?
Yes. POST /v1beta1/{+name}:watch performs a long-poll request that returns as soon as the named variable changes value or a server-side timeout elapses. This avoids client-side polling and gives near-real-time propagation of feature flags and rollout signals.
What are the rate limits for the Runtime Configuration API?
Google Cloud applies per-project read and write quotas to the Runtime Configurator, with watch and waiter calls counted against a separate concurrent-long-poll quota. The default per-minute quota is 600 read requests per project, raisable through the Cloud Console for sustained usage.
How do I create a waiter through Jentic?
Search Jentic for 'create runtime configurator waiter', load the schema for POST /v1beta1/{+parent}/waiters, and execute with the cardinality success condition and timeout you need. Jentic returns the waiter resource, which the agent polls until status reports done.
Is the Runtime Configuration API free?
The API is free to call within Google Cloud quotas, and stored variables and waiters incur no per-byte charge inside published limits. Costs scale only with very large variable bodies above the per-config size limits.
How do I scope variable access to one team only?
Place each team's variables in its own config and bind a team-specific IAM principal to the config with POST /v1beta1/{+resource}:setIamPolicy at the role roles/runtimeconfig.admin or .viewer. Other principals can be denied entirely at the config level rather than on each variable.
GET STARTED