Skip to content

ledger: reliability columns opt-out for publication-gate benches, dense table stays in its card - #2548

Merged
Flotapponnier merged 1 commit into
devfrom
fix/ledger-reliability-flag
Sep 20, 2026
Merged

Flotapponnier merged 1 commit into
devfrom
fix/ledger-reliability-flag

Conversation

@Flotapponnier

Copy link
Copy Markdown
Collaborator

On bench 268 the providers' success query is a publication gate (tfq_health, 0 or 1), so the ledger's Reliability block printed "100.00 %" on every published row and "0.00 %" with the RPC hint ("every call fails, the endpoint is still probed") on a product whose window is filling. With nine custom columns the dense layout also bled 80 to 128 px past its card at xl and 2xl, the header rules crossing the card border and the last columns clipped (Florent's screenshot, 2026-09-20).

A spec flag, ledger_reliability: false (default true, nothing changes elsewhere): the Reliability colgroup, the Success header and cells and the Errors column are not rendered, the empty-state and offline colspans follow, and the unresponsive pill reads "Filling" with a threshold hint. The dense wrapper keeps overflow-x-auto inside the card and drops the negative margins. Bench 268 opts in.

Deploy: worker rebuild + materialize reset (spec), staging Action (site).

🤖 Generated with Claude Code

https://claude.ai/code/session_01L4rjxemTvgRwrCG6D2iK5D

…olumns and labels an unpublished row Filling (threshold wording) instead of Unresponsive (probe wording); the dense layout stays inside its card instead of bleeding 80 to 128 px past it; bench 268 opts in

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01L4rjxemTvgRwrCG6D2iK5D
@Flotapponnier
Flotapponnier merged commit a14ef2b into dev Sep 20, 2026
1 of 2 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.

2 participants