From 0da6f90ad21a5c99e256cc2f155db735e64b125d Mon Sep 17 00:00:00 2001
From: Viktor Barzin <viktorbarzin@meta.com>
Date: Sun, 1 Mar 2026 14:44:22 +0000
Subject: [PATCH] =?UTF-8?q?[ci=20skip]=20openclaw:=20fix=20slow=20startup?=
 =?UTF-8?q?=20=E2=80=94=20proper=20resources=20+=20readiness=20probe=20+?=
 =?UTF-8?q?=20VPA=20off?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Set explicit CPU (2 cores) and memory (2Gi) limits
  Root cause: Goldilocks VPA was throttling to 300m CPU, causing gateway
  to take 5+ minutes to start, and 1Gi memory caused OOM crashes
- Add TCP readiness probe on port 18789 to prevent 502 Bad Gateway
  during startup (Traefik was routing before gateway was listening)
- Disable Goldilocks VPA via namespace label (vpa-update-mode: off)
---
 stacks/openclaw/main.tf | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/stacks/openclaw/main.tf b/stacks/openclaw/main.tf
index edcd500e..47d80b6e 100644
--- a/stacks/openclaw/main.tf
+++ b/stacks/openclaw/main.tf
@@ -13,7 +13,8 @@ resource "kubernetes_namespace" "openclaw" {
   metadata {
     name = "openclaw"
     labels = {
-      tier = local.tiers.aux
+      tier                                       = local.tiers.aux
+      "goldilocks.fairwinds.com/vpa-update-mode" = "off"
     }
   }
 }
@@ -358,6 +359,13 @@ resource "kubernetes_deployment" "openclaw" {
           port {
             container_port = 18789
           }
+          readiness_probe {
+            tcp_socket {
+              port = 18789
+            }
+            initial_delay_seconds = 30
+            period_seconds        = 10
+          }
           env {
             name  = "OPENCLAW_GATEWAY_TOKEN"
             value = random_password.gateway_token.result
@@ -432,10 +440,12 @@ resource "kubernetes_deployment" "openclaw" {
           }
           resources {
             limits = {
-              memory = "1Gi"
+              cpu    = "2"
+              memory = "2Gi"
             }
             requests = {
-              memory = "64Mi"
+              cpu    = "100m"
+              memory = "512Mi"
             }
           }
         }