From d50e1a925622664cdd7fd4883de84bc39c86402a Mon Sep 17 00:00:00 2001
From: Thomas Flament <thomas.flament@scality.com>
Date: Fri, 13 Mar 2026 12:29:34 +0100
Subject: [PATCH] Increase Kind apiserver event TTL from 1h to 4h
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Kubernetes events expire after 1 hour by default. In CI, kind export
logs often runs 1-2 hours after the test window, by which time the
relevant events have already been garbage-collected.

This was a concrete gap during the Azure archive flakiness
investigation (PR #2340): the failing tests ran around 06:36 but kind
export logs didn't execute until ~08:22 — nearly 2 hours later. All
Kubernetes events from the failure window (pod scheduling, restarts,
liveness probe failures, OOM kills) had expired and were unavailable
for diagnosis.

Bump --event-ttl to 4h via a kubeadm ClusterConfiguration patch in
the Kind cluster config, ensuring events survive for the entire
duration of CI jobs.

Issue: ZENKO-5217
---
 .github/scripts/end2end/bootstrap-kind.sh | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/.github/scripts/end2end/bootstrap-kind.sh b/.github/scripts/end2end/bootstrap-kind.sh
index 774eceb54e..3be0dbe465 100755
--- a/.github/scripts/end2end/bootstrap-kind.sh
+++ b/.github/scripts/end2end/bootstrap-kind.sh
@@ -61,6 +61,11 @@ nodes:
     nodeRegistration:
       kubeletExtraArgs:
         node-labels: "ingress-ready=true"
+  - |
+    kind: ClusterConfiguration
+    apiServer:
+      extraArgs:
+        event-ttl: "4h0m0s"
   extraMounts:
   - hostPath: ${VOLUME_ROOT}/data
     containerPath: /data