Skip to content

fix(website): show what the CLI really does, and restore the phone views - #605

Merged
bmscomp merged 1 commit into
mainfrom
fix/website-proof
Nov 5, 2026
Merged

bmscomp merged 1 commit into
mainfrom
fix/website-proof

Conversation

@bmscomp

@bmscomp bmscomp commented Nov 5, 2026

Copy link
Copy Markdown
Owner

Why

A design critique scored the site 18/28. The worst finding: the page's two main examples were refused by the CLI. kates resilience run --type LOAD --fault broker-pod-kill failed with unknown flag: --type, and kates test lint found four errors in the page's scenario.yaml. These are the critique's three P1s.

What changed

  • Scenario: the page shows the exact file of a real run, with that run's output, labelled real run, kates v1.37.0, Kind, 3 brokers, 2026-11-05. maxErrorRate is gone because kates test apply never checks it, and the copy no longer says a scenario injects a fault.
  • CI: website.yml now lints the scenario from the page, so it can't drift again.
  • Sample data: the hero terminal is labelled sample output, and the section states that the two figures are illustrative.
  • Accuracy: corrected the resilience command, --min-grade (now C everywhere), POD_KILL and the krafter-brokers-N names.
  • Quick start: dropped "five commands" and "in minutes", and added the git checkout, the current-context warning, make clean, and a note that the gate exits 1 below the grade.
  • Mobile: the vertical diagram's boxes have their names back, and Kates is the first comparison column.

Checks

  • html-validate@9 passes.
  • The page's scenario, extracted the way the new CI step does it, passes go run . test lint, and it matches the file that ran.
  • Browser at 375px: all six box names show, the broker labels fit, no sideways scroll.

Deliberately not done

  • A real broker kill. The lab runs chaos as noop, and enabling it needs a cluster change that was not made. A real run can replace the sample numbers later.
  • The real hero numbers. On Kind, one LOAD run gave 88.3K rec/s and a 546 ms P99, and kates gate --min-grade C failed at grade D, so the sample figures are far tidier than a laptop result.
  • An "on my cluster" path and a measured install time, because neither was tested.
  • The critique's P2s (page order and the closing ask, including a Sponsor link).

The comparison table's column order changes on desktop too.

The hero command and scenario.yaml could not run: resilience run has no --type
or --fault, and test lint found four unknown fields in the scenario. The
scenario is now the file of a real run, its output is that run's, and a new
step in the website workflow lints it so the page can't drift again. Its copy
no longer says a scenario injects a fault, and maxErrorRate goes because test
apply never checks it.

No fault ran on the lab, so the hero terminal is labelled sample output and the
two figures illustrative. Their commands, gate grade, POD_KILL and broker names
are corrected.

The quick start drops "five commands" and "in minutes", and adds the checkout,
the current-context warning and make clean. On phones the diagram's boxes get
their names back, and Kates moves to the comparison's first column.
@bmscomp
bmscomp merged commit 8659605 into main Nov 5, 2026
60 checks passed
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