The portal

The portal is the web app at cockpit.graphpilot.io. It is where you create organizations and services, invite your team, watch your traffic, read what a deployment built, purge the cache, and manage billing. Everything that changes what the edge serves (your schema.graphql and gpilot.toml) can also be deployed from here, but most teams deploy from the CLI or from CI and use the portal to look.

Getting an account

Sign up at cockpit.graphpilot.io/signup. After you sign in, the portal opens your organization. If you belong to several, the sign-in asks which one to continue with. If you belong to none yet, it opens the Organizations page, where you create one with New organization.

The same account signs you in to the CLI: gpilot auth login --browser opens the sign-in in your browser. See signing in to the CLI.

The sidebar changes with where you are. At the top of it, the organization switcher lists your organizations and, under each, its services, so you can jump from one service to another without going back to a list.

Organization

Entry	What it holds	Documented in
Dashboard	Every service in the organization as a card, the request allowance used this period, New service, and, until you dismiss it, the organization’s onboarding checklist. A service’s own dashboard has a checklist of its own.	Services and settings
Members	Who belongs to the organization, their roles, and invitations.	Organizations and members
Billing	Your plan, usage, payment method, invoices and billing account.	Billing and usage
Settings → General	The organization’s name, URL slug, website and industry, and deleting the organization.	Organizations and members
Settings → Domains	Domains the organization has verified, for custom hostnames.	Custom domains
Settings → API keys	Organization keys and personal keys.	API keys

Service

Entry	What it holds	Documented in
Dashboard	Requests, cache hit rate, errors and latency for the last 24 hours, open alerts and top operations.	Services and settings
Insights	Your request telemetry: Requests and Operations.	Insights
Alerts	Every time one of GraphPilot’s alert rules fired for this service.	Alerts
Deployments	What is live, what each deployment built and found, comparing two, and New deployment.	Deployments
Purges	The Purge button, and the history of every purge, automatic and manual.	Purges
Queries	Persisted queries (your manifest) and APQ (what clients registered).	Queries
Schema	The live schema: Explorer, Cache rules, Surrogate keys, Operations, Findings and schema.graphql.	Schema and config findings
Config	The live configuration: Values, Findings and gpilot.toml.	Schema and config findings
Settings	General (name, origin, signing key, origin allowlist, introspection, deleting the service), Domains, Alerts and Runtime.	Services and settings

A service that has never gone live shows a short getting-started panel on its Dashboard, Insights, Alerts, Purges, Queries and Schema pages instead of empty charts. It names the command to run next (gpilot deploy) and links to the quickstart.

Account

Open the menu under your name at the bottom of the sidebar. It lists the three account sections.

Section	What it holds	Documented in
Profile	Your personal details, and Delete account at the foot of the page.	Your account
Security	Password and multi-factor authentication.	Your account
Sessions	The browsers and devices signed in to your account.	Your account

Hidden or disabled

What you see depends on your permissions and your plan, and the portal shows them differently:

A control you lack the permission for is not shown. Without deployment:edit there is no New deployment button, and without cache:purge there are no purge buttons in Insights or on the Schema pages. Ask an admin of your organization for a role that carries it. Every permission is on permissions.
A page you may not read at all says so. Deployments without deployment:view, for example, says that you do not have permission.
A control your plan does not include is shown disabled, with the reason. Point at it to see which plan includes it, for example an Insights filter or column, a latency percentile, or New service once the plan’s services are in use.

A few controls are exceptions and say so where they are documented: the Purge button on the Purges page stays visible and disabled without cache:purge, because the page would read as broken without it, and Delete organization is disabled rather than hidden without organization:delete.

When something does not load

A page whose data cannot be loaded says so in place of its content, for example Could not load billing, with Try again.

When the page loads but a single value on it cannot be read, the rest of the page is shown and the value’s place reads This value could not be loaded., with Try again where there is room for it. The portal does not show a missing value as an answer: a usage figure that failed is not Currently unavailable, a payment method that failed is not No payment method (and offers no Add), and a requester that failed is not Not recorded or Automatic. In the request panel of Insights, the panel says Some details of this request could not be loaded. above its tabs.

A value that decides what you can do is never shown as missing. When the plan, the account lock, a quota, a signing key, a runtime pin, a deployment’s schema or the name typed to confirm a deletion cannot be read, the page shows its error instead. Insights figures, charts and tables also fail as a whole rather than show a partial answer.

A value already on screen stays when a later refresh cannot read it again.

Getting help from the portal

The help button in the bottom corner of every page, and the Help and feedback group in the account menu, open the documentation, the FAQ, Contact support and Send feedback. See contacting support.

Quickstart: create a service and deploy it
Services and settings
Insights
Deployments and Purges