From 1f80be5cce8f0254c6778030aa0865fb28e11bf8 Mon Sep 17 00:00:00 2001 From: npub1qyvc0c5kl4gqv2fd97fsk46tu378sqgy35vc83rvgfwne90sel7s0ed67d <011987e296fd5006292d2f930b574be47c7801048d1983c46c425d3c95f0cffd@buzz.block.builderlab.xyz> Date: Thu, 30 Jul 2026 14:09:18 -0400 Subject: [PATCH] docs: re-anchor problem-statement clock to the fitted post-onset window MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Dawn's delta review at 3292b35db caught the doc carrying two runways: the problem statement still held kickoff's eyeballed ~3.5-4d doubling / 8-10d saturation, while the summary row read the fitted ~7d-to-80% — and the eyeball figure was fitted across exactly the Jul-21-18:00Z regime boundary the adjacent fit-window bullet forbids. Problem statement now leads with the fitted post-onset clock (~3.4-3.5d doubling, ~7d to 80%, r2=0.92) and demotes the kickoff figures to a declared pre-onset eyeball. Same treatment for the ~150x/~400x family growth (fitted: x9.6-12.4 over 9 days). G0's gate description (line ~595) keeps its original tasking text deliberately. Model and configs untouched; doc only. Co-authored-by: npub1qyvc0c5kl4gqv2fd97fsk46tu378sqgy35vc83rvgfwne90sel7s0ed67d <011987e296fd5006292d2f930b574be47c7801048d1983c46c425d3c95f0cffd@buzz.block.builderlab.xyz> Signed-off-by: npub1qyvc0c5kl4gqv2fd97fsk46tu378sqgy35vc83rvgfwne90sel7s0ed67d <011987e296fd5006292d2f930b574be47c7801048d1983c46c425d3c95f0cffd@buzz.block.builderlab.xyz> --- docs/redis-cluster-mode.md | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/docs/redis-cluster-mode.md b/docs/redis-cluster-mode.md index b321c1d695..cda7cb0eb6 100644 --- a/docs/redis-cluster-mode.md +++ b/docs/redis-cluster-mode.md @@ -74,11 +74,16 @@ disabled, auto-failover **enabled**, Multi-AZ **enabled** (verified via `describe-replication-groups` / `describe-cache-clusters`, profile `bb-public-operations-ro`). -- `EngineCPUUtilization` (primary): 0.3% (Jul 21) → 8.4% (Jul 27) → 19.2% - (Jul 30) — doubling every ~3.5–4 days. Saturation in ~8–10 days if the - curve holds. -- Driver is command volume: get/set/eval families each ~150x in two weeks, - pub/sub ~400x. Product growth (~80k users added), not a leak. +- `EngineCPUUtilization` (primary): 0.3% (Jul 21) → 8.4% (Jul 27) → 19.5% + (Jul 30). **Fitted post-onset clock (P3, log-linear, r²=0.92): doubling + every ~3.4–3.5 days; ~7 days to 80%, ~8 to saturation.** (Kickoff's + eyeballed "~3.5–4 days / 8–10 days" spanned the pre-onset regime — see the + fit-window bullet below — and understated the urgency; the fitted number + is the citable one.) +- Driver is command volume — product growth (~80k users added), not a leak. + Fitted 9-day post-onset growth is x9.6–12.4 across all command families + (kickoff's "~150x / ~400x in two weeks" were coarser eyeballs over a + window spanning the onset). - Whole-instance CPU ~4%, memory 0.09%, replication lag <1ms. Vertical scaling is dead: same-family bigger instances have identical single-core speed; r8g (Graviton4) buys ~25–30% single-thread ≈ 1.5 days at the current