Until now anyone admitted through Dex saw every namespace and every log, read with the module's service account. auth.visibility: ViewerRBAC (the default) asks the API server as the viewer, with the token the DexAuthenticator forwards: list pods in all namespaces means the whole cluster; otherwise the namespaces where they can list pods. Logs also need get pods/log, node cards list nodes, cluster-wide alerts are for cluster-wide viewers. Traffic to a hidden namespace keeps its counts with the peer shown as hidden. One minute cache per token hash, SSARs in parallel (8), a cluster-wide viewer costs two. Verified against the live cluster with two real tokens: a service account with view in one namespace got that namespace only (0 nodes, 403 on another namespace's YAML and logs, 403 on node usage); the module's own account, cluster-wide, got all 56 namespaces and 6 nodes; no token, 401. auth.visibility: Everything keeps the old behaviour. The DexAuthenticator now forwards the token when either actions or ViewerRBAC need it (templates/_helpers.tpl). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
52 lines
4.2 KiB
YAML
52 lines
4.2 KiB
YAML
v0.10.0:
|
|
features:
|
|
- "Per-user visibility (auth.visibility: ViewerRBAC, the default): each viewer sees the namespaces where their own RBAC lets them list pods, or the whole cluster if they can list pods everywhere. Logs also need get pods/log, node cards list nodes; traffic peers in hidden namespaces are shown as hidden. Checked with SelfSubjectAccessReviews as the viewer, cached for a minute. auth.visibility: Everything keeps the previous behaviour."
|
|
v0.9.0:
|
|
features:
|
|
- "Vulnerabilities from operator-trivy: a red shield with the number of CRITICAL CVEs on each card, and in the details every CRITICAL and HIGH CVE with its packages, installed and fixed versions and a link. A CVE counts once per card."
|
|
fixes:
|
|
- "Event counts are scaled to the 30-minute window: a cumulative count spread over days (40 probe failures in two days) no longer badges a card as a burst."
|
|
- upmeter's smoke-mini services and pods, which it moves between nodes on purpose, are not reported as problems.
|
|
v0.8.3:
|
|
fixes:
|
|
- UnexpectedJob warnings (the CronJob controller's note on every manual run) are listed for information and no longer badge the CronJob.
|
|
v0.8.2:
|
|
fixes:
|
|
- An Ingress that answers with a redirect (redirect annotations, or rewrite ... permanent / return 30x in a snippet) is not reported for a backend Service that does not exist; Deckhouse's prometheus Ingress does exactly that.
|
|
- Warning events about a pod that no longer exists are listed for information and no longer badge the card.
|
|
v0.8.1:
|
|
fixes:
|
|
- "Actions: the DexAuthenticator is annotated allow-access-to-kubernetes, so the user's token carries the kubernetes audience and the API server accepts it. In v0.8.0 the token lacked that audience, the API server rejected it and actions were unavailable."
|
|
v0.8.0:
|
|
features:
|
|
- "Problem badges: a yellow or red \"!\" on cards, namespace and node boxes and in the tree, from object state (crash loops, pending pods, readiness, failed CronJob runs, Services without endpoints, node pressure), Warning events and Deckhouse ClusterAlerts."
|
|
- A Problems panel with every current problem, most severe first, filterable to applications or system namespaces.
|
|
- "Actions from the details panel: restart, scale, run a CronJob now, suspend and resume, delete a pod, cordon and uncordon a node. They run with the viewer's own token and RBAC; the module's service account stays read-only."
|
|
- "Load section: CPU and memory of a workload or node against requests, limits and allocatable, with a per-pod table."
|
|
- An Events tab with the events of the object, its pods and ReplicaSets.
|
|
- Cards are locked in place until the padlock is opened; the map can be saved as a PNG picture.
|
|
- The open page refreshes cards, badges and the tree every minute without moving the layout.
|
|
fixes:
|
|
- "Layout of a large namespace or node view is 3-4 times faster: levels without edges are packed without a call into ELK."
|
|
- "About 20% fewer flows through hubble-relay: the to-overlay copy of pod-to-pod traffic is filtered at the agents; traffic counts are unchanged."
|
|
- Static files are revalidated with ETags instead of downloaded again on every visit (about 2 MB).
|
|
v0.7.2:
|
|
fixes:
|
|
- A namespace without traffic in the selected window is drawn again; the empty window was sent as null and the map kept the previous namespace on screen.
|
|
- A map that fails to draw shows an error instead of the previous view.
|
|
v0.7.1:
|
|
fixes:
|
|
- Alert rules are rendered on DKP versions without the operator-prometheus-crd module.
|
|
v0.7.0:
|
|
features:
|
|
- First release as a Deckhouse module (ModuleSource, release channels, lib-helm templates).
|
|
- "Traffic collection modes: Continuous, OnDemand, Disabled; retention configurable up to 24h."
|
|
- Hubble CA is read by a module hook from d8-cni-cilium; no manual copy of the certificate.
|
|
- Authentication through user-authn (DexAuthenticator) or any external auth, as in cilium-hubble.
|
|
- English and Russian UI.
|
|
- Prometheus metrics and alerts for the Hubble stream.
|
|
fixes:
|
|
- Traffic windows now cover the requested time instead of the agent ring buffer (about 10 seconds).
|
|
- Counts are distinct connections, not Hubble events.
|
|
- Pod exits cleanly on SIGTERM.
|