From 25c3d4f688e20eaf7e53b17cddc0b95844a8e538 Mon Sep 17 00:00:00 2001
From: Graham McIntire <graham@mcintire.me>
Date: Sun, 19 Apr 2026 16:35:16 -0500
Subject: [PATCH] fix(grid-rs): bump memory limit to 1 Gi
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The real worker binary OOM-killed at 512 Mi on startup. The plan's
512 Mi assumption was speculative — actual Rust RSS at boot includes
sqlx pool buffers, reqwest connection pool, tokio runtime, and
tracing JSON subscriber. 1 Gi leaves room for all of that plus the
steady-state ~500 Mi during a chain step.
---
 k8s/deployment-grid-rs.yaml | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/k8s/deployment-grid-rs.yaml b/k8s/deployment-grid-rs.yaml
index c7560490..d27532a2 100644
--- a/k8s/deployment-grid-rs.yaml
+++ b/k8s/deployment-grid-rs.yaml
@@ -71,10 +71,14 @@ spec:
           resources:
             requests:
               cpu: 100m
-              memory: 128Mi
+              memory: 256Mi
             limits:
               cpu: "4"
-              memory: 512Mi
+              # 512 Mi OOM-killed the real binary at startup. 1 Gi gives
+              # room for sqlx pool + reqwest pool + tokio runtime +
+              # tracing buffers, with headroom vs the steady-state
+              # target (~500 Mi during a chain step).
+              memory: 1Gi
           securityContext:
             allowPrivilegeEscalation: false
             runAsNonRoot: true