The dashboard

doze status is a snapshot. Sometimes you want the movie — to watch a service wake up, see connections come and go, tail logs, or actually reach into a queue and send a message without leaving the terminal. That’s doze dash.

doze dash

It opens a live TUI (built with Charm’s Bubble Tea — lovely toolkit, worth a look if you build terminal apps) that reflects the daemon’s real state and updates as things change. If the daemon isn’t running, dash starts it for you.

What you’re looking at

The main view is your stack, grouped by category — databases, caches, queues, storage, your own process blocks — each service on a line with its state, endpoint, connection count, and live resource use. The state is the honest picture of doze’s whole lazy-boot idea, in one column:

• reaped — asleep, costing nothing. Its endpoint is still there; the next connection wakes it.
• booting — someone just connected (or you asked); it’s coming up.
• active — running, with connections.
• idle — running, but nothing’s talking to it. In a few minutes it’ll reap itself. This is the state you’ll watch drift to reaped after you close your app, which is oddly satisfying.
• tainted — something went wrong on boot; the details are one keypress away so you’re never guessing.

Arrow keys move around, tab switches focus, and there’s a filter if your stack is big. Select a service and you can wake or sleep it right there, or drop into its logs.

The AWS console, built in

Here’s the part that’s more than a status screen. Select one of the local AWS engines — S3, SQS, SNS — and the dashboard becomes a little data console for it. No aws CLI, no separate GUI, no hunting for the right endpoint flag.

• S3 — browse your buckets and objects, and put an object straight from the dashboard to test an upload path.
• SQS — watch queues fill and drain, send a message (FIFO-aware, so it’ll ask for the bits it needs), inspect what’s sitting in a queue, and purge when you want a clean slate. The dead-letter and redrive setup is visible too, so you can actually see a message get dead-lettered instead of guessing why it vanished.
• SNS — see your topics and subscriptions, and publish a message with attributes so you can watch it fan out to the subscribing queues — filter policies and all. It’s the fastest way to answer “is my subscription filter actually matching?” without writing a throwaway script.

This is the kind of thing that usually means alt-tabbing to a web console or copy-pasting an awslocal incantation. Having it a keystroke away, pointed at your running stack, is a small thing that saves a surprising number of little context switches.

Getting around

The footer always shows the keys for wherever you are, so you don’t have to memorize anything — but the shape of it:

• ↑ ↓ / arrows — move within a list
• tab — switch between panels
• enter — open / drill into the selected thing
• esc — back out
• / — filter
• action keys (send, publish, put, purge, wake, sleep…) shown in context
• ? — help · q — quit

When to reach for it

Honestly? Whenever status isn’t enough. When you’re debugging a message that should’ve been delivered and wasn’t, when you want to watch a slow first boot finish, when you’re eyeballing memory across the stack, or when you just want to poke a queue and see what happens. It’s not required for anything — every action in it has a command-line equivalent — but for the exploratory, “what’s actually going on” moments, a live view you can steer beats a snapshot you have to keep re-running.