Repository navigation
feat(session-environments): polish the health card and name the bound environment on a locked chip - #27480
Conversation
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: these are copy, colour and display-choice changes on a card that has screenshots. The locked-chip fix makes the gateway's lock decide which environment the chip names. [DESIGN-REVIEWED] d2f1cbb Description read[DESCRIPTION-READ] d54ce46226f92d28440ed58b564b04fd41f8fcde90ecf7fa8c7432c3c37a7c7b That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Every item comes from the #25241 follow-up list. The only extra surface is an optional Notes
What this change shipsInventory (8 items) — 7 justifiedIntent: a user can read the remote workspace's health at a glance and always sees which machine a locked chat runs on. ADDITION (polish) plus one FIX.
[FIRST-PRINCIPLES-REVIEWED] d2f1cbb Description read[DESCRIPTION-READ] d6dd4871467e6423dc918a54e8e95287aaa20a474c5e9f970683463ca0762b1d That sha256 names the PR description this verdict read. Recompute it with |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing blocks: the blind reader identified every relabelled row, the neutral meters and the low-memory line correctly, and the card has no controls to misread. Notes
[UX-REVIEWED] d2f1cbb Description read[DESCRIPTION-READ] 6cdcb6a07d2ee214d85443201594d8bf997fc2308b3b0f15009ce353e7a3e3e0 That sha256 names everything this verdict read: the PR description, plus the 7 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
d1417d0 to
e4c3161
Compare
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
… environment on a locked chip
e4c3161 to
d2f1cbb
Compare
|
Rebased: main moved in website/src/i18n/locales/ (bn, de, en-XA, en.manual, es, fr, hi, it, ja, ko, pt, ru, zh-CN) and website/src/components/ since this head went green (f2c0d92 -> 763378e); CI re-running. No conflicts; the chip, locale and i18n tests pass on the new head d2f1cbb. 🤖 AI-assisted comment 🤖 |
|
Review-ready at head d2f1cbb.
Closes #25241. No merge, approval or auto-merge from me. 🤖 AI-assisted comment 🤖 |
CrysisDeu
left a comment
There was a problem hiding this comment.
Approve at d2f1cbb7. 5 new keys in all 13 catalogs with real translations; 3 removed keys have no users. The locked chip shows the profile name from config, React-escaped; the health card shows real API state with no default "healthy". CI green, no overrides.
Nits: three UX-lane notes still open; one unused fallback icon; branch is behind main, rebase before merge.
Posted by code-reviewer (Kiro Crew member) on behalf of @CrysisDeu. Full report on file.
bolichen97
left a comment
There was a problem hiding this comment.
Reviewed at d2f1cbb. No blocking findings; LGTM.
Non-blocking notes (follow-up is fine):
- ebsite/src/components/EnvironmentHealthCard.tsx:~73 -
iconis optional with a<Server>fallback, but the only caller always passes it. For an unknown provider,ProviderGlyphalready falls back to a generic cloud, so theServerbranch is unreachable. Makeiconrequired or drop the fallback (the first-principl - ebsite/src/i18n/locales/en.manual.json:~10245 - "Busy cores" overstates what the number is. A Linux load average also counts tasks in uninterruptible I/O wait, so a value of 3 "of 4" does not mean three cores are busy. A label like "Load vs cores" is more accurate. The UX review also found the three-number "x · y · z
- ebsite/src/i18n/locales/en.manual.json:~10247 - "Memory is low. Heavy builds may fail." gives no next step, and the 11px
text-warnline reads weakly in light theme (UX review notes). Optional copy and contrast follow-up, not a defect.
Problem / Motivation
Goal: The remote environment's health card reads plainly, and a locked chip always names the environment the chat is bound to.
The follow-up list on #25241, left over from #21259:
Why it matters
The health card is how a user checks whether a remote workspace can take a heavy build. Labels that need decoding, and a chip that names the wrong machine, make the user guess where their commands run.
Not a goal
What changed (motivation → approach → change)
Each item from the issue is a small, separate change to the two components and the copy.
0.42 · 0.3 · 0.250.42 · 0.3 · 0.25 of 4bg-muted); only warn and danger are colouredEnvironmentHealthCardtakes anicon)lockedis trueThe locked-chip fix is one line in
SessionEnvironmentChip.tsx: once the gateway says the chat is locked, a pick still in flight can only be refused, so the chip usesdata.profileinstead of the pending target. Before the lock, a pending pick still shows at once, as today.The three relabelled strings get new keys (
statusHeading,workspaceName,busyCores), so no catalog keeps an old translation of the old wording. Two keys are new (busyCoresValue,memoryLow). All 12 hand-written catalogs inwebsite/src/i18n/locales/carry them, anden-XAis regenerated.test/test_coder_session_launch.pygets a comment on its"stats"refusal: the guest browse has astatsop, and the Files browse route still cannot reach it, which is what the oldtest_the_browse_mode_has_no_metrics_oppin guarded.Design calls, each the most conservative option:
·separator.Backwards compatibility
Compatible: only display copy, colour and one label choice change; no API, config or stored data is touched.
Reader: website/src/components/EnvironmentHealthCard.tsx:EnvironmentHealthCard -- crew page -- the only reader of
sessionEnvironments.health.readOnly,.workspaceand.load; it reads the renamed keys (SessionEnvironmentChip.test.tsxasserts the new copy)Tests
website/src/test/SessionEnvironmentChip.test.tsx:shows the remote environment metrics when hovering the fixed chipasserts the new heading, row label, busy-cores label and0.1 · 0.2 · 0.3 of 2value, the Coder mark in the card, a neutral healthy meter, and no advice line.advises on low memory and colours only a warning meter(new): at 85% used, the advice line shows, the memory meter isbg-warnand the healthy disk meter isbg-muted.keeps a remote chat on its health card, refusal shown, when the lock overtakes a switch to This machineasserts that the chip reads "Coder MicroVM" with the Coder mark, never "This machine", while the refused switch is pending.All three fail with the component changes reverted and pass with them (40/40 in the file). Locale style, parity and pseudolocale tests pass with
I18N_BASE_REFset;tsc,eslint,i18n:checkandi18n:renderpass.Manual verification
Captured the real built dashboard with stubbed APIs: hovered a chat locked to a Coder profile, in dark and light themes, with healthy memory and with memory at 88% used.
Screenshots / video
Healthy (dark, light):
Low memory (dark, light):
Related Issues
Closes #25241
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)