[bong-water-water-bong]

Two cells of the #315 record were never measured, and the gaps table gave them a reason but no result, which reads as
"measured, nothing to report". This marks both explicitly and replaces the stale reasons.

cell	before	after
GLM throughput at the soak composition	"the box went down before the llama-bench run"	not measured … "not yet measured - needs one llama-bench hold" plus the exact command (llama-bench -m GLM-4.7-Flash-Q4_K_M.gguf -dev HRX0 -ngl 99 -fa on -p 512,2048 -n 128 -r 3)
the forced-eviction intervention cell (~/wt/evictonly) re-run with the fix	"…needs the services-off window the owner gates …, and the box is down"	not measured — the #315 regime needs ~100 GiB of GTT with four GLM-4.7-Flash neighbours, no services-off window has opened, "so this cell has no result and none is implied" → "not yet measured - needs the owner's services-off window"

Both reasons had gone stale in a way that mattered: the box has been up since the reboot, and the GPU is held by the
leaderboard and kernel lanes, not by a crash. Marking them unmeasured is the honest state, and it keeps the two cells
from being read as passing evidence for a closed entry.

Also corrected in the same file: the entry's closing status still said the documentation was "waiting in PR #518"; it is
on main.

Docs only — no code, no behaviour change, KNOWN-ISSUES.md alone.

🤖 Generated with Claude Code

[github-actions]

PR Reviewer Guide 🔍

(Review updated until commit 7e79ca4)

Here are some key observations to aid the review process:

🎫 Ticket compliance analysis 🔶 315 - Partially compliant Compliant requirements: The PR correctly documents the resolution of the issue through the soak testing The PR updates the documentation to reflect that the issue is resolved The PR correctly identifies and marks the two previously unmeasured cells as "not yet measured" Non-compliant requirements: None Requires further human verification: None

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Documentation Clarity The PR updates the documentation to clarify that two cells were never measured, but the change also introduces a slight inconsistency in how the status is described. The original description said the documentation was "waiting in PR #518", but the PR now says it's on main. This is a minor discrepancy in the documentation update. Status of the documentation itself: this entry and the docs change are in `KNOWN-ISSUES.md` on main (they landed in #518); #315 is closed on the strength of the soak above.

[github-actions]

Persistent review updated to latest commit 7e79ca4