The test is implemented as a go test.
It can be executed as a bazel target.
The general workflow is to setup a cluster,
point the test to the workspace in which to
find the kubeconfig and the constellation config
and specify a target image, k8s and
service version. The test will succeed
if it detects all target versions in the cluster
within the configured timeout.
The CI automates the above steps.
A separate workflow is introduced as there
are multiple input fields to the test.
Adding all of these to the manual e2e test
seemed confusing.
Co-authored-by: Fabian Kammel <fk@edgeless.systems>
* bazel: add configuration for remote caching
* ci: enable bazel remote caching for building binaries
* ci: use bazel directly when building go binaries
* ci: enable cache for most build steps
* dev-docs: document remote caching
K-Bench's network benchmarks require two distinct worker nodes.
Add check prior to running the benchmark that terminates early, if not
enough workers scheduled.
Add the option to run K-Bench performance to the manual CI workflow
Install CSI drivers in the cluster for K-Bench benchmarks
Attach the results to the workflow in the GitHub Actions view
* e2e tests now execute on the latest debug image available by default
* e2e-manual workflow now takes an optional image reference to run on
* isDebugImage is a flag that has to be set in case
you are running a debug image
* e2e-test workflows execute two hours earlier.
* Run quick-mode e2e tests for the two older versions we support.
This triggers every night, together with the existing e2e tests.
Idea here is that we know that a cluster can be setup and initialized.
* Run full e2e tests for the two older versions each sunday.
* Do not abort manual e2e runs. This allows for parallel runs.
* Run unprivileged container
* cleaned up actions and new measure action to generate, sign and upload measurements
* improve constellation ip fetching to support multiple control nodes
Signed-off-by: Fabian Kammel <fk@edgeless.systems>
* Make e2e pipeline use the latest image available.
* Use pcr-reader to read & store measurements.
* buildvcs false in ci
* only notify teams on main
* plain yq syntax, since if already checks for csp
* previous version of yq requires explicit eval
* fix pcr-reader call
* actually pass variable between jobs
* fix typo
* Make order of images consistent.
* read measurements after create
Co-authored-by: Fabian Kammel <fk@edgelss.systems>
