From 6b97d01861ab55794b808bc20e7a014265ad5ff4 Mon Sep 17 00:00:00 2001 From: tegwick Date: Mon, 21 Sep 2026 18:36:31 +0200 Subject: [PATCH] Restore State Hub's CPU request to the chart's 250m on railiance01. The 100m override worked around a node with <250m unallocated. Knative and Kourier released ~830m of requests on 2026-09-21, and State Hub bursts past 400m under concurrent agent load while reserving only 100m. Co-Authored-By: Claude Opus 5 Assistant: claude-code Assistant-Model: opus Assistant-Process: 63291@bnt-lap001 Assistant-Session: 8bd77868-ca68-4f49-bb1e-d539ecc0d703 --- deploy/railiance/apps/helm/state-hub-values.yaml | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/deploy/railiance/apps/helm/state-hub-values.yaml b/deploy/railiance/apps/helm/state-hub-values.yaml index 3a464a5..6bebcb0 100644 --- a/deploy/railiance/apps/helm/state-hub-values.yaml +++ b/deploy/railiance/apps/helm/state-hub-values.yaml @@ -35,10 +35,12 @@ config: sbomNexusWriteMode: nexus resources: - # The single 4-core node currently has less than 250m unallocated. Keep enough - # headroom for maxSurge=1 so State Hub can roll without an outage. + # Restored to the chart's 250m on 2026-09-21. The earlier 100m was a workaround + # for a node with <250m unallocated; ~830m of Knative/Kourier requests were + # released that day (the-custodian docs/kubernetes-change-gate-decision.md). + # State Hub idles near 5m but bursts past 400m under concurrent agent load. requests: - cpu: 100m + cpu: 250m memory: 512Mi limits: cpu: 1000m