From a33eb6369c404ab912e4124086e855ffa4f60e1e Mon Sep 17 00:00:00 2001 From: Yuan Gao Date: Thu, 13 Aug 2026 11:55:50 -0700 Subject: [PATCH] e2e: note that not every failed suite retains a namespace The "After a failure" section promises a kept namespace, but retention only covers namespaces the suite created. A suite running against a pre-installed worker pool holds its atespace in Valkey and leaves nothing labeled ate.dev/e2e behind, so point the reader at the standing demo namespace. --- internal/e2e/README.md | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/internal/e2e/README.md b/internal/e2e/README.md index cf6b200f3a..63a209c278 100644 --- a/internal/e2e/README.md +++ b/internal/e2e/README.md @@ -40,6 +40,16 @@ of running pods, so clean up once you are done reading: $ hack/cleanup-e2e.sh # deletes every namespace labeled ate.dev/e2e ``` +Not every suite has a namespace to keep. One that runs against a pre-installed +worker pool -- `TestActorEgress` uses the standing `ate-demo-egress` pool from +`hack/install-ate-kind.sh --deploy-demo-egress` -- creates only an atespace, +which lives in Valkey rather than Kubernetes, so a failure leaves nothing +labeled `ate.dev/e2e` behind. Read the standing namespace instead: + +```shell +$ kubectl logs -n ate-demo-egress -l ate.dev/worker-pool=egress +``` + ## Creating a new test suite Copy `testmain_test.go` from `internal/e2e/suites/example` into your new suite. It will