From fb236e18f71a8aeb3276675d6a608b457161aa70 Mon Sep 17 00:00:00 2001 From: viniciusdc Date: Tue, 16 Jun 2026 09:53:00 -0300 Subject: [PATCH] ci: cap the E2E job with timeout-minutes so hung setup can't squat a runner The E2E job had no timeout, so when a setup step hangs (a foundational-services install recently wedged on an unbounded readiness wait, well before the operator is even deployed) the job sat in-progress until GitHub's 6h default, holding a runner the whole time. Add a 30-minute job cap -- a healthy run finishes well under that, so this only bites genuine hangs, failing them in minutes. This is a standalone resilience fix. Replacing the hand-rolled foundational setup (dev/scripts/services/*) with the nebari-sandbox action is tracked separately. --- .github/workflows/build-pr.yml | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/.github/workflows/build-pr.yml b/.github/workflows/build-pr.yml index 791ab97..906a503 100644 --- a/.github/workflows/build-pr.yml +++ b/.github/workflows/build-pr.yml @@ -99,6 +99,10 @@ jobs: name: E2E Tests runs-on: ubuntu-latest needs: lint + # Hard cap so a hung setup step (e.g. an unbounded wait while installing + # foundational services) fails the job in minutes instead of squatting a + # runner until GitHub's 6h default. A healthy run finishes well under this. + timeout-minutes: 30 env: CLUSTER_NAME: test-cluster steps: