Flow Studio

On the [Flow Studio] page, you can see how your service actually flows — from account connection through applications, events, and cost — on one canvas. Every other menu shows you its own list; this page shows the connections between them, so you can spot where the flow broke and jump straight to the menu that fixes it.

Page Layout

The page has three layers, stacked top to bottom.

Selecting a feature on the map opens a detail panel on the right. The map is a read-and-navigate surface: you inspect a feature here, then open the owning menu to make the change.

Reading the Map

Sections, features, and tasks

The map nests three levels, and the words are worth keeping straight because the screen uses them consistently:

Features that are not part of the required path are grouped under Optional setup, which has no section-run button.

Node states

Each feature node shows its name, a status, and a short measure of what it currently holds.

A greyed-out node means there is nothing for you to register — it either runs once the earlier feature is in place, or its data lands from outside.

A node that is waiting says what it is waiting for in words, naming the feature: Create 「Cloud Account」 and this runs, or Once 「Resource Collection」 finishes, this runs.

The measure on a node

The number under a node’s name is labelled in the cell itself — 128 resources, 12 rules, 43 of 50 accounts, 3 running — rather than being a bare figure you have to recognise. Where a feature works through targets one by one, the node shows progress against the total, and how many of them failed or are still going.

Select the ⓘ on a node to read what that particular measure counts.

Expanding a feature

A feature that runs more than one job can be expanded to show the tasks behind it, either from the node itself or from Execution breakdown in the detail panel.

Features with nothing to expand say so — a single save or a read-only view has no tasks behind it.

Connections

Select the information icon at the bottom-left of the canvas to open Reading the map, which explains what the lines mean.

Choosing What the Map Shows

The status row carries two controls. They look alike, but they act on the map in different ways, and the difference is deliberate.

Views

The view switch decides which features the map draws at all.

Switching away from Everything removes the other features from the canvas rather than fading them, so the map genuinely gets smaller. A connection that only ran through a removed feature is folded up, joining the remaining features directly, so the flow still reads end to end instead of looking broken. Features that remain keep the position they had, so nothing has to be found again.

Status chips

The chips beside the feature count — Set up, Running jobs, Failed, Warning, Waiting, Not set up, State unknown — pick out the features in one state. Hovering a chip explains what it lights up.

Unlike the view switch, a chip dims everything else instead of removing it. The chip answers which features are in this state, and the rest of the map has to stay on screen for that answer to be readable as a part of the whole. The map also moves to the matching features, so you do not have to look for them.

What Is Running Right Now

When runs are under way, an In progress band sits above the map and says what the map alone cannot: that work is happening, on what, and since when.

Each entry names the feature and reads out:

[Open ] jumps the map to the feature that is running.

The band does not refresh itself — it is as of the Checked time, same as the map. Use [Refresh] to see where a run has got to.

Feature Detail

Selecting a feature opens its detail panel. Every feature uses the same five sections, so the panel reads the same way wherever you are.

The button at the top of the panel opens the menu that owns the results — [Open Resource Inventory] in the example above. It opens in a new tab so you keep the map. Features whose data arrives from an external pipeline have no such button.

Runs by

The Runs by row tells you what a run covers, which is not the same for every feature.

Features that need attention

When a feature has open issues, the panel leads with them — an issue count next to the status badge and a callout above the summary.

The callout is written for the case at hand rather than being a generic warning:

Where several accounts failed for different reasons, it says so and sends you into the feature rather than guessing a single cause.

Running

Running a section

Selecting [Run section] on a section header runs that whole section and opens the detail panel of the first feature that actually runs, so you can watch it start.

Where the section can be narrowed, the button offers both: [Run all] takes the whole scope, [Run selected] opens the scope picker below. The button also states the scope it is about to take — Run the Resources section · 12 accounts, Run the Cost section · whole tenant — so you can read what will happen before selecting it.

A section with nothing runnable in it says Nothing to run here rather than offering a button that would do nothing.

Choosing what a run covers

[Run selected] opens a dialog listing what the run can be narrowed to. It comes in two forms, depending on what is being run.

Most sections open Run Accounts — a table of accounts with their provider, access type, credential status, and collection status, and a count of what you have selected. Collection opens the fuller form below, which adds two more steps.

There is a cap on how many accounts one run can take. Past it the dialog selects the first N and tells you the limit rather than silently truncating.

Running a single feature

The detail panel’s Running section runs the feature on its own. Features that follow it will not run — and where those are known, the panel names them.

Long-running features estimate the wait (About 12 min) and, past a threshold, suggest you start it and leave the page.

If the flow definition changed while you had the page open, the run is not started against the stale version — the page reloads the current definition and asks you to run it again.

Run History

Select [Run history] in the header to open the run-history screen. Use the back arrow next to the title to return to the map.

Use the buttons on the right to expand or collapse every row at once.

Results

Narrowing the list

Two rows of controls sit above the list, and they work on different axes.

Tabs — All and Mine — split by who ran it. Filters — Filter by section, Period (last 24 hours / 7 days / 30 days), and Result — narrow which runs. [Clear filters] resets the filter row.

Task results

Expanding a row shows the result of each task in that run, plus a re-run button.

Each task reports Succeeded, Failed, Needs a look, Running, or Waiting, with counts where it worked through targets — 8 of 50 targets failed. [Show technical details] reveals the raw error for a failure whose cause could not be named.

The button matches what there is to re-run: [Re-run this part] for a section, [Re-run N accounts] where the run covered accounts, and [Run this step again] for a standalone feature run. While that section is still running, the button is disabled until it finishes.

Related

Connecting a cloud account and collecting its resources are the first features of this flow. To register or repair an account, see Cloud Account.