Fleet
For administrators who need the state of every Studio deployment at once. Viewers can read the list; creating a deployment needs the DeploymentOperator Role or the Owner or Operator tier.
Search and filter
Open Fleet from the dock or the Fleet Management group. The search bar matches the start of a deployment's app domain, its deployment ID, or its account's alias or account ID, and is written to the address bar as ?q=, so a filtered view can be bookmarked. It does not match the Name column on its own: to find a deployment by name, type the start of its app domain, its account alias or its ID. The address also accepts ?account= and ?status= to narrow to one AWS account or one health value; there are no on-screen controls for those two, and an unknown status is ignored.
Start a new deployment
New deployment opens the deployment wizard (Deploy your first Studio). Without the permission it is disabled with the hover "Requires the DeploymentOperator role (or owner/operator)".
Read the columns
| Column | What it shows |
|---|---|
| Name | The name you gave the deployment, with its app domain beneath (a deployment created by an older release shows its domain-derived name). A governed deployment adds " · Governed" (hover: everyone assigned sees every agent, workflow, connector and knowledge graph). " (stale)" means health was last checked more than 15 minutes ago. |
| Account | Alias and id, or the id alone. |
| Region | The AWS region. |
| Network access | Private, or Public (CloudFront, existing) for a deployment created before private network access. |
| Outbound traffic | NAT gateway or Transit Gateway (hover: the full wording, "Through AlphaAgent's NAT gateway" or "Through your Transit Gateway (your firewall)"). |
| VPC range | The deployment's VPC range; 10.60.0.0/16 for a deployment created before the range was asked. |
| App domain | The Studio hostname, or a dash before one is recorded. |
| Version | The Studio release (wording below). |
| Status | The health value (below); while an install job runs the hover reads "Job <id> in progress". |
| Licence | The state the deployment last reported: active, inactive, or a dash. |
| Last job | "Waiting for you — resolve", opening the exact parked step (for example accepting a Transit Gateway attachment), when a job waits on you; otherwise a link to the running or last job. |
Clicking a row opens the deployment detail. The list refreshes every 30 seconds while the tab is visible and pages with Previous and Next, 20 rows at a time; the search runs on the server across the whole fleet (Lists, search and paging).
Address plan
Address plan, in the Fleet Management group, lists every VPC range this Organisation holds: "Private deployments are routed on your network, so each range must be unique across your address plan: a new deployment's range may not overlap any range listed here, or a VPC already in its AWS account." The first row is the Organisation's own range; then one row per deployment, sorted by range, with VPC range, Held by (the deployment), Account, Region, Outbound traffic, Subnets and State. A removed deployment keeps its range until its VPC is gone and Verify on its account confirms it (hover: "This range stays held until the VPC is gone. Once it is, verify the account under Accounts to release the range."). Clicking a deployment row opens its detail.
Read Status and Version
Status is Provisioning, Awaiting input, Healthy, Degraded, Failed, Deleting or Disabled (paused); while an install job runs it shows Installing… instead of Provisioning or Degraded. A private deployment reads Degraded when the health check finds its Transit Gateway attachment or routes broken. Once a deployment is uninstalled it leaves the Fleet list; its own page stays readable at its address and reads Uninstalled, and Audit keeps its history.
Version reads the running version when nothing is changing. Otherwise: "<recorded> → updating to <target>" during an update job; "Not yet running → <release>" before the first task starts; "<running> (recorded <release>)" when the health poll saw a newer task definition than the last recorded release (the record catches up on the next upgrade); "Paused at <release>" or "Paused"; "Uninstalling…", "Uninstall failed: <reason>" or "Uninstalled"; a dash when no version is known.
Empty and error states: No deployments yet ("Register an AWS account, then start a new deployment to install AlphaAgent Studio into it.", with New deployment if you may create one); No deployment matches these filters; The fleet could not be read with "HTTP <status>. The rows below may be out of date." and Retry for a server-side failure; or the permission message from The console: layout and roles.
What you should see
- Every installed deployment on one row, with its network access, outbound traffic and VPC range; the address bar carrying your search as
?q=. - A row waiting on you showing Waiting for you: resolve in Last job.
Notes
- The list does not create, pause, update or delete deployments; open a row for those.
- The Licence column shows the state the deployment reported, not the Console's licence page (Console: licences, usage and billing).
- A "(stale)" row is not necessarily broken; its detail page reads live health.
- "Uninstall failed: …" is explained on the deployment's Jobs view (Jobs, progress and failures).