ankra stack-profiles demo or the API.
Before you start
- The organisation has a staging cluster, and its agent is online. Every demo lands there. Someone who manages Ankra AI sets it under Ankra AI → AI workspaces, in Staging cluster, or with
ankra org ai-environment set --staging-cluster <cluster>. See AI Workspaces. - You can read the stack profile. Your own profiles, profiles shared with your organisation, and public profiles can all be demoed. The demo runs on your staging cluster, so demoing someone else’s public profile uses none of their resources.
- Your role holds
stacks.deploy. A demo is a real deployment through the resource engine, and launching or stopping one needs that permission. - The Ankra CLI is installed and signed in. See Ankra CLI.
Launch and follow a demo
1
Launch the demo
Give the profile’s name or id. Without flags, the demo runs the profile’s current version with its default values, for the organisation’s default lifetime:Pin a version with Ankra creates the namespace, applies its quota, and deploys the profile into it with every member forced into that one namespace. The command prints the response as JSON, with the demo’s
--version, set the lifetime with --ttl-hours, and bind parameters with --set name=value. For a secret parameter, use --set-env name=ENV_VAR or --set-file name=path, which keep the value out of your shell history:workspace_id, its namespace and the version it runs. The commands below take the workspace id.2
Check that it is listed
demos, beside your lifetime policy and the staging cluster.3
Watch it come up
inspection holds the steps (namespace created, guardrails applied, stack deployed), the resources the profile installed, and the namespace’s Kubernetes events. A profile demo deploys no workload of its own, so there is no image or routing step to wait on: the stack operation completing is its readiness. If a chart fails to install, the failure text is the stack’s own, not a generic demo failure.4
Read its logs
The demo has no pod yet, so there are no logs to read.One demo per profile
A profile has at most one live demo per organisation. Its namespace is derived from the organisation and profile ids,ankra-demo-sp-<org>-<profile> with the first eight hexadecimal characters of each, so launching again replaces the running demo rather than adding a second copy. That is also why renaming a profile never strands a demo.
Launch again to pick up a new profile version or changed parameter values.
Lifetime
Without--ttl-hours, a demo lives for the organisation’s demo lifetime, 24 hours unless someone changed it (ankra org ai-environment set --demo-ttl-hours). Your role sets the longest lifetime you may ask for; a longer one is refused with Your role allows demos of at most <n> hours.
Capacity
The demo namespace carries a hardResourceQuota at the full-stack tier, because a profile installs a whole stack rather than one workload:
Stop a demo
ankra stack-profiles demo list my-profile again: once the teardown has finished, the demo is no longer listed.
Where else profile demos are used
- Previews built from a profile. An application’s preview settings can name a stack profile (Choose a profile), so every preview of that application starts its own copy of the profile beside the app: see Set what every preview starts with. When you only want to look at the stack, a profile demo is cheaper than a whole preview.
- Publishing a managed service. When you publish a managed service package over a stack profile, Ankra can fill the package’s connection from a running demo of that profile and try it there before you publish: see Publish a managed service.
What can go wrong
API
All endpoints sit under the profile and require authentication; the mutations require a CSRF header for browser-originated requests. Bearer-token twins live under/api/v1/org/stack-profiles/... for the CLI.
See Stack Profiles for authoring and versioning, and Preview a branch for the application lane.