Fix flaky test by disabling Prometheus metrics #1341
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The test suites for some of our packages spin up parts of a Kubernetes
cluster using the `envtest` library, so that we have something more
solid than mocks to test against. One problem this was causing was that
running tests for both the `kube` and `server` packages at once would
cause one of those test suites to fail.
Side-note: it's possible to reproduce this problem a little more
efficiently by spinning up a fake web server to sit on port 8080, for
example using Python's built-in web server:
python3 -m http.server 8080
After some digging, I found out that the reason for this is that both
environments were trying to spin up a Prometheus metrics handler
listening on port 8080. If they both tried to do this at once, one would
fail, causing the entire test suite for that package to fail too
(including some nil-pointer panics during the clean up, which might be
worth tidying up in a future pull request).
Setting the metrics bind address to the string "0" disables metrics for
the test clusters (the documentation for which is buried in a comment in
the Kubernetes docs[1]), which fixes this particular problem. Testing on
my local machine, there didn't seem to be any other fixed port
assignments cropping up to take its place.
[1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager
dhwthompson
force-pushed
the
fix-double-test-port-failure
branch
from
59d39fa
to
c6076a2
Compare
jpellizzari
approved these changes
Jan 25, 2022
J-Thompson12
approved these changes
Jan 25, 2022
joshri
pushed a commit
that referenced
this pull request
Jan 26, 2022
The test suites for some of our packages spin up parts of a Kubernetes
cluster using the `envtest` library, so that we have something more
solid than mocks to test against. One problem this was causing was that
running tests for both the `kube` and `server` packages at once would
cause one of those test suites to fail.
Side-note: it's possible to reproduce this problem a little more
efficiently by spinning up a fake web server to sit on port 8080, for
example using Python's built-in web server:
python3 -m http.server 8080
After some digging, I found out that the reason for this is that both
environments were trying to spin up a Prometheus metrics handler
listening on port 8080. If they both tried to do this at once, one would
fail, causing the entire test suite for that package to fail too
(including some nil-pointer panics during the clean up, which might be
worth tidying up in a future pull request).
Setting the metrics bind address to the string "0" disables metrics for
the test clusters (the documentation for which is buried in a comment in
the Kubernetes docs[1]), which fixes this particular problem. Testing on
my local machine, there didn't seem to be any other fixed port
assignments cropping up to take its place.
[1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager
jpellizzari
pushed a commit
that referenced
this pull request
Jan 26, 2022
The test suites for some of our packages spin up parts of a Kubernetes
cluster using the `envtest` library, so that we have something more
solid than mocks to test against. One problem this was causing was that
running tests for both the `kube` and `server` packages at once would
cause one of those test suites to fail.
Side-note: it's possible to reproduce this problem a little more
efficiently by spinning up a fake web server to sit on port 8080, for
example using Python's built-in web server:
python3 -m http.server 8080
After some digging, I found out that the reason for this is that both
environments were trying to spin up a Prometheus metrics handler
listening on port 8080. If they both tried to do this at once, one would
fail, causing the entire test suite for that package to fail too
(including some nil-pointer panics during the clean up, which might be
worth tidying up in a future pull request).
Setting the metrics bind address to the string "0" disables metrics for
the test clusters (the documentation for which is buried in a comment in
the Kubernetes docs[1]), which fixes this particular problem. Testing on
my local machine, there didn't seem to be any other fixed port
assignments cropping up to take its place.
[1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager
jpellizzari
pushed a commit
that referenced
this pull request
Jan 26, 2022
The test suites for some of our packages spin up parts of a Kubernetes
cluster using the `envtest` library, so that we have something more
solid than mocks to test against. One problem this was causing was that
running tests for both the `kube` and `server` packages at once would
cause one of those test suites to fail.
Side-note: it's possible to reproduce this problem a little more
efficiently by spinning up a fake web server to sit on port 8080, for
example using Python's built-in web server:
python3 -m http.server 8080
After some digging, I found out that the reason for this is that both
environments were trying to spin up a Prometheus metrics handler
listening on port 8080. If they both tried to do this at once, one would
fail, causing the entire test suite for that package to fail too
(including some nil-pointer panics during the clean up, which might be
worth tidying up in a future pull request).
Setting the metrics bind address to the string "0" disables metrics for
the test clusters (the documentation for which is buried in a comment in
the Kubernetes docs[1]), which fixes this particular problem. Testing on
my local machine, there didn't seem to be any other fixed port
assignments cropping up to take its place.
[1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager
jpellizzari
added a commit
that referenced
this pull request
Jan 28, 2022
* New proto endpoints * Added integration tests * V2 workflow added * Use controller-runtime client in app server * Fix whitepace and unused linting errors * Add kustomization tests * Fix flaky test by disabling Prometheus metrics (#1341) The test suites for some of our packages spin up parts of a Kubernetes cluster using the `envtest` library, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both the `kube` and `server` packages at once would cause one of those test suites to fail. Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server: python3 -m http.server 8080 After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request). Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs[1]), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place. [1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager * Use suite test for kubebuilder setup/teardown * Add more create kustomization test coverage * Skip acceptance test lint * Split k8s resources into separate files * Initial work implementing HelmRepository to the API * Added weave related labels to helm repository resource * Use app name to list helm repositories * Added unit test for ListHelmRepositories * Added missing namespace parameter when listing helm repositories Co-authored-by: James Wilson <jwils22@gmail.com> Co-authored-by: Jordan Pellizzari <jordan@weave.works> Co-authored-by: David Thompson <32775+dhwthompson@users.noreply.github.com>
jpellizzari
added a commit
that referenced
this pull request
Feb 7, 2022
* New proto endpoints * Added integration tests * V2 workflow added * Use controller-runtime client in app server * Fix whitepace and unused linting errors * Add kustomization tests * Fix flaky test by disabling Prometheus metrics (#1341) The test suites for some of our packages spin up parts of a Kubernetes cluster using the `envtest` library, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both the `kube` and `server` packages at once would cause one of those test suites to fail. Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server: python3 -m http.server 8080 After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request). Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs[1]), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place. [1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager * Use suite test for kubebuilder setup/teardown * Add more create kustomization test coverage * Skip acceptance test lint * Split k8s resources into separate files * Initial work implementing HelmRepository to the API * Added weave related labels to helm repository resource * Use app name to list helm repositories * Added unit test for ListHelmRepositories * Added missing namespace parameter when listing helm repositories Co-authored-by: James Wilson <jwils22@gmail.com> Co-authored-by: Jordan Pellizzari <jordan@weave.works> Co-authored-by: David Thompson <32775+dhwthompson@users.noreply.github.com>
joshri
pushed a commit
that referenced
this pull request
Feb 10, 2022
* New proto endpoints * Added integration tests * V2 workflow added * Use controller-runtime client in app server * Fix whitepace and unused linting errors * Add kustomization tests * Fix flaky test by disabling Prometheus metrics (#1341) The test suites for some of our packages spin up parts of a Kubernetes cluster using the `envtest` library, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both the `kube` and `server` packages at once would cause one of those test suites to fail. Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server: python3 -m http.server 8080 After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request). Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs[1]), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place. [1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager * Use suite test for kubebuilder setup/teardown * Add more create kustomization test coverage * Skip acceptance test lint * Split k8s resources into separate files * Initial work implementing HelmRepository to the API * Added weave related labels to helm repository resource * Use app name to list helm repositories * Added unit test for ListHelmRepositories * Added missing namespace parameter when listing helm repositories Co-authored-by: James Wilson <jwils22@gmail.com> Co-authored-by: Jordan Pellizzari <jordan@weave.works> Co-authored-by: David Thompson <32775+dhwthompson@users.noreply.github.com>
joshri
pushed a commit
that referenced
this pull request
Feb 16, 2022
* New proto endpoints * Added integration tests * V2 workflow added * Use controller-runtime client in app server * Fix whitepace and unused linting errors * Add kustomization tests * Fix flaky test by disabling Prometheus metrics (#1341) The test suites for some of our packages spin up parts of a Kubernetes cluster using the `envtest` library, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both the `kube` and `server` packages at once would cause one of those test suites to fail. Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server: python3 -m http.server 8080 After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request). Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs[1]), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place. [1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager * Use suite test for kubebuilder setup/teardown * Add more create kustomization test coverage * Skip acceptance test lint * Split k8s resources into separate files * Initial work implementing HelmRepository to the API * Added weave related labels to helm repository resource * Use app name to list helm repositories * Added unit test for ListHelmRepositories * Added missing namespace parameter when listing helm repositories Co-authored-by: James Wilson <jwils22@gmail.com> Co-authored-by: Jordan Pellizzari <jordan@weave.works> Co-authored-by: David Thompson <32775+dhwthompson@users.noreply.github.com>
jpellizzari
added a commit
that referenced
this pull request
Feb 25, 2022
* New proto endpoints * Added integration tests * V2 workflow added * Use controller-runtime client in app server * Fix whitepace and unused linting errors * Add kustomization tests * Fix flaky test by disabling Prometheus metrics (#1341) The test suites for some of our packages spin up parts of a Kubernetes cluster using the `envtest` library, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both the `kube` and `server` packages at once would cause one of those test suites to fail. Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server: python3 -m http.server 8080 After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request). Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs[1]), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place. [1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager * Use suite test for kubebuilder setup/teardown * Add more create kustomization test coverage * Skip acceptance test lint * Split k8s resources into separate files * Initial work implementing HelmRepository to the API * Added weave related labels to helm repository resource * Use app name to list helm repositories * Added unit test for ListHelmRepositories * Added missing namespace parameter when listing helm repositories Co-authored-by: James Wilson <jwils22@gmail.com> Co-authored-by: Jordan Pellizzari <jordan@weave.works> Co-authored-by: David Thompson <32775+dhwthompson@users.noreply.github.com>
jpellizzari
added a commit
that referenced
this pull request
Mar 3, 2022
* New proto endpoints * Added integration tests * V2 workflow added * Use controller-runtime client in app server * Fix whitepace and unused linting errors * Add kustomization tests * Fix flaky test by disabling Prometheus metrics (#1341) The test suites for some of our packages spin up parts of a Kubernetes cluster using the `envtest` library, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both the `kube` and `server` packages at once would cause one of those test suites to fail. Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server: python3 -m http.server 8080 After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request). Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs[1]), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place. [1]: https://pkg.go.dev/sigs.k8s.io/controller-runtime@v0.10.2/pkg/manager * Use suite test for kubebuilder setup/teardown * Add more create kustomization test coverage * Skip acceptance test lint * Split k8s resources into separate files * Initial work implementing HelmRepository to the API * Added weave related labels to helm repository resource * Use app name to list helm repositories * Added unit test for ListHelmRepositories * Added missing namespace parameter when listing helm repositories Co-authored-by: James Wilson <jwils22@gmail.com> Co-authored-by: Jordan Pellizzari <jordan@weave.works> Co-authored-by: David Thompson <32775+dhwthompson@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The test suites for some of our packages spin up parts of a Kubernetes cluster using the
envtestlibrary, so that we have something more solid than mocks to test against. One problem this was causing was that running tests for both thekubeandserverpackages at once would cause one of those test suites to fail.Side-note: it's possible to reproduce this problem a little more efficiently by spinning up a fake web server to sit on port 8080, for example using Python's built-in web server:
After some digging, I found out that the reason for this is that both environments were trying to spin up a Prometheus metrics handler listening on port 8080. If they both tried to do this at once, one would fail, causing the entire test suite for that package to fail too (including some nil-pointer panics during the clean up, which might be worth tidying up in a future pull request).
Setting the metrics bind address to the string "0" disables metrics for the test clusters (the documentation for which is buried in a comment in the Kubernetes docs), which fixes this particular problem. Testing on my local machine, there didn't seem to be any other fixed port assignments cropping up to take its place.
Closes: #1332