Repository navigation
Close the production-readiness gaps - #36
Merged
Merged
Conversation
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Three gaps kept "production-ready out of the box" from being true for an in-cluster install. This closes them.
Reads need a credential. A server that holds a bootstrap token, single sign-on, or any API key now answers 401 to an /api request that carries neither a bearer token nor a session cookie. Before, those reads were served as an anonymous viewer, so anyone who could reach the server could read every report. A server with no credential at all, --insecure, and --demo keep anonymous reads. A key created with the CLI against the same store closes anonymous reads within ten seconds. Health, readiness, the OpenAPI spec, the OIDC paths, signed inbound webhooks, and the dashboard's static files stay open. The dashboard shows a sign-in panel on a 401 and offers the token drawer, and the live event stream now reads through fetch so it carries the bearer token.
The chart can scan a fleet. kubeconfig.existingSecret mounts a kubeconfig Secret read-only and passes it as --kubeconfig. A pod given no kubeconfig used to find zero contexts, so scheduled scans did nothing. It now scans its own cluster through the ServiceAccount as the context in-cluster. The ClusterScan controller stays on the host cluster when a fleet kubeconfig is mounted, unless --controller-context names a context.
The image is published. The release workflow gains packages: write, QEMU, buildx, and a ghcr.io login, and stops skipping docker, so each release pushes signed amd64 and arm64 images and their manifests. The chart moves to 0.15.0.
Tests: TestAnonymousReadsOnlyWithoutCredentials, TestHoldsCredentialSeesLateKey, TestDashboardSignIn (browser), TestContextsOrInCluster, TestControllerKubeconfig, and TestChartKubeconfigSecret, each shown to fail with its fix reverted. Six tests that asserted anonymous reads on a token server now assert 401. A local goreleaser snapshot built both images, and the arm64 image served 401 to an anonymous read and 200 with the token.