Quickstart
Usually dev:up brings it up
Section titled “Usually dev:up brings it up”On its own, dashboard shows nothing. It needs something to read (the ledger’s API and the
development issuer) and the daemon that answers validation and contents. The dev:up of
capture-ledger (the ledger that decides which
URLs get captured and records what came back) starts all of them, and the dashboard as part of it.
cd ../capture-ledgerpnpm run dev:up # prints the dashboard's URL when it is doneopen http://127.0.0.1:7080/dev:up expects this repository next to capture-ledger (../dashboard). If you keep it
somewhere else, use pnpm run dev:up --dashboard <path>. How the whole stack comes up is in
capture-ledger’s quickstart.
| To | Run in capture-ledger |
|---|---|
| restart only the dashboard | pnpm run dev:up --from dashboard |
| see what is running | pnpm run dev:status |
| stop everything | pnpm run dev:down (also stops what runs on the host) |
When you only work on the dashboard
Section titled “When you only work on the dashboard”If you only change the dashboard, start it from this repository. It has no npm dependencies, so there is nothing to install.
npm run dev # http://127.0.0.1:7080/The ledger’s API and the issuer have to be up (the ones dev:up started are enough). 検証
(validate) and the contents view also need the daemon. If you would rather not start the ledger’s
stack, open it in the sandbox instead:
npm run sandbox starts a stand-in for the ledger and the real daemon in containers, and serves
the dashboard on 7081.
Changes under public/ take effect when you reload the browser, because the server reads the
files from disk on every request. After changing src/, restart it. If the dashboard that dev:up
started already holds 7080, npm run dev says so and exits (see the table below). To run both
side by side, move yours to another port (DASHBOARD_PORT=7081 npm run dev).
Grant your identity access
Section titled “Grant your identity access”By default the dashboard identifies itself to the ledger with this machine’s user name
(DASHBOARD_SUBJECT in Configuration). That name needs permission to submit
crawls for the organization.
cd ../capture-ledgerpnpm run fga:grant submitter "$(whoami)" acme # onceIf you followed capture-ledger’s quickstart, this is already done. Without it, the top of the
screen says not found in red and the lists stay empty. For an identity without the
permission, the ledger returns an empty list of crawls and a 404 for the scripts, and the
dashboard stops at that 404 without drawing the lists.
Fix it from the message
Section titled “Fix it from the message”The dashboard itself is in Japanese, so the messages below are quoted as they appear.
| Where | Message | What to do |
|---|---|---|
| top of the screen (red) | not found |
The identity has no permission. Run the fga:grant above |
| top of the screen (red) | … に届かない (…)。capture-ledger で pnpm run oidc:issuer は動いているか? |
The development issuer is not up. Start it with dev:up |
| top of the screen (red) | … が名乗りを断った (…)。… |
The issuer refused the identity. Look at the issuer’s terminal |
| top of the screen (red) | … に届かない (…)。capture-ledger で pnpm run api は動いているか? |
The ledger’s API is not up. Start it with dev:up |
| validation cell, contents view | … に届かない (…)。capture-ledger で pnpm run dev:up --from validator は済んでいるか? |
The daemon is not up. Run the command it names |
| identity line | dashboard 自身に届きません |
The dashboard’s own server has stopped. Start it again |
| footer | 取りに行けませんでした（上の表は最後に描けたもの） |
The reason is the red message at the top. The tables are the last ones that could be drawn |
| scripts view (red) | 台帳にスクリプトがありません。… |
The ledger has no scripts to run. Import them with the command in the message |
| new crawl view, under 始める | 404 —— not found |
The identity may not start crawls. Run the fga:grant above |
| new crawl view, under 始める | 404 —— Route POST:/api/crawls not found |
The ledger does not know where to send crawls (Windmill). The dev:up stack has it |
| crawls view (it moves there) | the running row is marked | A 409: only one crawl runs at a time, overall. Wait for the marked row to finish |
| terminal | 127.0.0.1:7080 は、既に動いている dashboard です (pid …)。 |
Just open it. To restart it, pnpm run dev:up --from dashboard |
| terminal | 127.0.0.1:7080 は別のものが握っています … |
Find out what it is with pnpm run dev:status, clean up with pnpm run dev:down |
When something fails, the dashboard keeps its tables. An empty table would read as “there is nothing to see”, which you could not tell apart from “it could not be fetched”.