Skip to content

Refresh the Studio screenshots after the Monitoring and instance rebuilds - #82

Open
rnv812 wants to merge 4 commits into
developfrom
docs/studio-screenshots
Open

Refresh the Studio screenshots after the Monitoring and instance rebuilds#82
rnv812 wants to merge 4 commits into
developfrom
docs/studio-screenshots

Conversation

@rnv812

@rnv812 rnv812 commented Aug 28, 2026

Copy link
Copy Markdown
Member

Closes #74.

Every screenshot on the site was taken before the rebuilds of Monitoring and the instance page, the identity plate of Management, the log channels, project import and export, and the Repositories entry the left rail has gained since — so every full-page shot documented an interface that is no longer there.

Screenshots

All 33 screens are shot again, light and dark, at 1920x1080 — 27 under studio_ui/ and 6 under home/, the last of which the landing page and the feature tour read.

The scope is wider than the issue's table on purpose: the left rail gained Repositories, and it is in every full-page shot, so a selective reshoot would have left the rest visibly older than the pages beside them. scenario_* were stale for a second reason the issue does not name — the data flow diagram is built from scripts as well as templates now — and repositories_* predated Discover and the paging under it by a few hours.

They come from one staged instance, so the screens agree with each other: eighteen projects, sixteen instances across every state, two scenarios exchanging four global keys through templates and a script event plugin, a repository connected with a generator installed from it, six secrets in the keyring, and an hour of generation behind the charts and the failure counts.

The release post

The 2.7.0 post illustrated itself from studio_ui/, which this reshoot rewrites. Its two images are now its own, under public/images/blog/eventum-2-7-0/, the way the 2.4.0 post already keeps them — a release post stays a record of that release.

Prose

A stale sentence is worse than a stale image, so the pages were read against the instance rather than only re-illustrated:

  • Monitoring — rewritten. It opens on one line of state instead of a row of stage totals, holds a window of 2.5, 10 or 30 minutes, and gathers the running instances into one section: a load chart split by instance or by stage, a search and quick filters that narrow the chart and the table together, and a row that opens what that instance is doing beside it. The per-instance table gained Output and Failures.
  • Instances — the Overview tab leads with what the run costs (processor, memory in queues, bytes in and out), and keeps its throughput, its pipeline, and what the instance is in one column each; About and Scenarios are one panel now.
  • Home — Explore leads to Repositories in place of the link that left Studio for the Hub.
  • Settings — the application log is split into Main, Server, Access and MCP channels, Logging carries the third-party level, and the events queue can be bounded in memory.

pnpm build passes.

Out of scope

  • The README hero in eventum-generator/eventum is the same shot; it ships in a PR of that repository.
  • The Shared state panel of the template console reads its scope once when the view mounts and is never refreshed after a run, so it reads 0 keys while the API already holds them. The screenshot was taken with its own refresh pressed; the fix belongs in its own issue.

rnv812 added 3 commits August 28, 2026 03:35
The 2.7.0 post illustrated itself from public/images/studio_ui, which the
documentation reshoots whenever Studio changes. Its two images move under
public/images/blog/eventum-2-7-0, where the release keeps them as the record
of what shipped.
Every screenshot on the site dated from before the rebuilds of Monitoring and
the instance page, the identity plate of Management, the log channels, project
import and export, and the Repositories section the left rail has gained since
- so every full page shot was of an interface no longer there.

All 33 screens are shot again, light and dark, from one instance: a workspace
of eighteen projects, sixteen instances across every state, two scenarios
exchanging global state through templates and a script, a repository connected
and a generator installed from it, and an hour of generation behind the
charts.
Monitoring opens on one line of state rather than a row of stage totals, holds
a window of 2.5, 10 or 30 minutes, and gathers the running instances into one
section - a load chart split by instance or by stage, quick filters and a
search that narrow it with the table, and a row that opens the details of that
instance. The instance overview leads with what the run costs, and keeps its
throughput, its pipeline and what the instance is in one column each. Home
sends you to Repositories in place of the Hub, the application log is split
into channels, and the events queue can be bounded in memory.
Read against the shots they replaced, seven of them said less than before:
the repository was closed over its catalog, the inspector stood on a stage
with one plugin instead of the four the page describes, the formatter had
nothing for its errors panel to report, an instance carried no parameters,
the queue estimate and the data flow diagram had scrolled out of frame, and
a scenario had lost a member.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant