From 6e1cd2a932bc005274bb41377e1be8c9b84cabf0 Mon Sep 17 00:00:00 2001 From: HARPER Jon Date: Tue, 10 Feb 2026 09:48:49 +0100 Subject: [PATCH] increase gateway memory (Xmx128m limit512Mi) -> (Xmx186m limit768Mi) In production it's the only pod with a steady rate of full GCs (up to 10 every 5 minutes) A weird aspect to investigate is that we even have a steady rate of full GCs at night when there is virtually no traffic. This steady state is like so: - Tenured Gen 67Mb - Eden space oscillating between 17Mb and 27mb every 2 minutes - survivor spaces 0 Maybe it's normal to allocate 10Mb of memory every 2 minutes even without traffic (only kubernetes probes, and some other internal periodic handlers??), so the problem is just that the oldgen is overflowing and everything has to be handled with full GCs. To investigate. No need for docker compose dev env because only one user --- k8s/resources/common/gateway-deployment.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/k8s/resources/common/gateway-deployment.yaml b/k8s/resources/common/gateway-deployment.yaml index dcec538a..418f3d49 100644 --- a/k8s/resources/common/gateway-deployment.yaml +++ b/k8s/resources/common/gateway-deployment.yaml @@ -7,7 +7,7 @@ metadata: version: "1" app.kubernetes.io/component: gridsuite-springboot annotations: - gridsuite.org/memory: memory-xxs + gridsuite.org/memory: memory-xs spec: selector: matchLabels: