tohjustin / kube-lineage Goto Github PK

View Code? Open in Web Editor NEW

384.0 8.0 24.0 963 KB

A CLI tool to display all dependencies or dependents of an object in a Kubernetes cluster.

License: Apache License 2.0

Go 92.59% Makefile 1.23% Shell 6.18%

kubernetes k8s kubectl kubectl-plugin kubectl-plugins helm dependency dependent

kube-lineage's People

Contributors

Stargazers

Watchers

kube-lineage's Issues

Query for multiple root elements by label

Thanks so much for your great work here!!!

Use Case

It'd be so amazing if there were the ability to query by label for "multiple root elements". For example say we have multiple objects all correlated by label, but with their own lineage's below them. They are related and I'd like to view them all at the same time, with a single query.

Rough proposed API

kubectl lineage --selector app=my-awesome-app as an example
Shorthand being -l

Rough result

Query for elements that don't have owner-refs and define them as "roots" and apply the standard algorithm and render from there.

Feature request: PVs and PVCs

Thank you for this wonderful tool!

I saw that it lists PVs and PVCs as supported object relationship, but it doesn't seem to make any connection between PVs and PVCs (doesn't seem to follow claimRefs). This would be great to be able to connect PVs with the rest of the graph (including pods, etc.)!

docs: dependencies vs dependents

Thank your very much for this great tool.

Could you please explain in the README (and/or in -h) the difference between dependencies and dependents?

Are you sure "dependencies" is the correct term?

❯ k lineage cluster host-cluster -o=wide
NAME                                                             READY   STATUS   AGE   RELATIONSHIPS
Cluster/host-cluster                                             True             69d   []
├── HCloudMachineTemplate/host-cluster-control-plane             -                69d   [OwnerReference]
├── HCloudMachineTemplate/host-cluster-md-0                      -                69d   [OwnerReference]
├── HetznerCluster/host-cluster                                  -                69d   [ControllerReference OwnerReference]
│   ├── Secret/hetzner                                           -                69d   [OwnerReference]
│   └── Secret/host-cluster-kubeconfig                           -                69d   [OwnerReference]

With option -D

❯ k lineage -D Secret/host-cluster-kubeconfig
NAME                                                 READY   STATUS   AGE
Secret/host-cluster-kubeconfig                       -                69d
├── HetznerCluster/host-cluster                      -                69d
│   └── Cluster/host-cluster                         True             69d
└── KubeadmControlPlane/host-cluster-control-plane   True             69d
    └── Cluster/host-cluster                         True             69d

For me -D looks like reverse-dependencies to me.

Column Ready shows "-" for some CRDs

I see this output for my CRDs:

NAME                                                                             READY   STATUS                                                                                                                                                                                                                                                                                           AGE
ClusterStrong/clusterstrong                                                        -                                                                                                                                                                                                                                                                                                        6h15m
├── ClusterStrongRelease/hello-apalla-1-26-v6                                      True                                                                                                                                                                                                                                                                                                     6h15m
└── HelloClusterStrongRelease/hello-apalla-1-26-v6                                 -                                                                                                                                                                                                                                                                                                        6h15m
ClusterStrong/clusterstrong                                                        -                                                                                                                                                                                                                                                                                                        6h15m
├── ClusterStrongRelease/hello-apalla-1-26-v6                                      True                                                                                                                                                                                                                                                                                                     6h15m
└── HelloClusterStrongRelease/hello-apalla-1-26-v6                                 -                                                                                                                                                                                                                                                                                                        6h15m
    ├── HelloNodeImageRelease/hello-apalla-1-26-controlplaneamd64cloud-v5          -                                                                                                                                                                                                                                                                                                        6h15m
    └── HelloNodeImageRelease/hello-apalla-1-26-workeramd64cloud-v5                -                                                                                                                                                            ly fetched new release tag hello-apalla-1-26-v6                                                                                                                                                                                                                  29m
    ├── HelloNodeImageRelease/hello-apalla-1-26-controlplaneamd64cloud-v5          -                                                                                                                                                                                                                                                                                                        6h15m
    └── HelloNodeImageRelease/hello-apalla-1-26-workeramd64cloud-v5                -

For some CRDs the Ready column is "True", but for some I see only "-".

If I use kubectl describe I see that the Status.Ready is true (a non optional boolean).

Why does linage show "-" instead of "True"?

Feature request: break text if too long.

I have an event with a long text from time to time:

│   ├── Event/test-dfkhje-7wnxg.177a0a977f4c6837                                       -       ControlPlaneUnhealthy: Waiting for control plane to pass preflight checks to continue reconciliation: Machine test-dfkhje-7wnxg-6snkf reports EtcdMemberHealthy condition is unknown (Failed to connect to the etcd pod on the test-dfkhje-control-plane-qhfkf-sszhr node: could not establish a connection to any etcd node: unable to create etcd client: context deadline exceeded) (x11)   5m15s

Since this means the whole column has this width, it has this drawback.

I can see it without line breaks by using a tiny font on the console (ctrl minus):

But as soon as I use a bigger font, the lines are ugly, because the age (last column) gets wrapped to the next line:

Feature request: break text if too long.

A lot of cves due to old code

I scanned it in my cluster and a lot of cves came up in it

I removed them and also added it to build standalone binaries in the git action for amd and arm

#14

I have my own fork if anyone needs it till its merged:
https://github.com/Avi-Robusta/kube-lineage/releases/tag/v2.0.2

Objects not discovered as dependent, but parent-object discovered as dependency

Hey @tohjustin 👋 :)
I think I've experienced a bug and hope this report helps improving your wonderful tool.
Please let me know if you want more info or something else!

Story time

I've installed package-operator which brings its own CRDs which are strictly organized in a hierarchy.

After installation of the operator [1], there is a tree of objects, glued together by ownerReferences that looks like this:

ClusterPackage/package-operator
-> ClusterObjectDeployment/package-operator
  -> ClusterObjectSet/package-operator-$hash
    -> Actual kubernetes objects

But some object relations are missing from kubectl-lineage's output. This is the output for kubectl lineage clusterpackage package-operator:

NAME                                                                                 READY   STATUS   AGE
ClusterPackage/package-operator                                                      -                3m35s
└── ClusterObjectDeployment/package-operator                                         -                3m34s
    └── ClusterObjectSet/package-operator-56496479f6                                 -                3m33s
        ├── ClusterRole/package-operator-remote-phase-manager                        -                3m17s
        ├── ClusterRoleBinding/package-operator                                      -                3m44s
        ├── CustomResourceDefinition/clusterobjectdeployments.package-operator.run   -                3m37s
        ├── CustomResourceDefinition/clusterobjectsetphases.package-operator.run     -                3m37s
        ├── CustomResourceDefinition/clusterobjectsets.package-operator.run          -                3m37s
        ├── CustomResourceDefinition/clusterobjectslice.package-operator.run         -                3m37s
        ├── CustomResourceDefinition/clusterobjecttemplates.package-operator.run     -                3m37s
        ├── CustomResourceDefinition/clusterpackages.package-operator.run            -                3m37s
        ├── CustomResourceDefinition/objectdeployments.package-operator.run          -                3m37s
        ├── CustomResourceDefinition/objectsetphases.package-operator.run            -                3m37s
        ├── CustomResourceDefinition/objectsets.package-operator.run                 -                3m37s
        ├── CustomResourceDefinition/objectslice.package-operator.run                -                3m37s
        ├── CustomResourceDefinition/objecttemplates.package-operator.run            -                3m37s
        ├── CustomResourceDefinition/packages.package-operator.run                   -                3m37s
        └── Namespace/package-operator-system                                        -                4m22s

I know that there is a missing ServiceAccount which I can reverse query for its dependencies with kubectl lineage -D -n package-operator-system ServiceAccount/package-operator:

NAMESPACE                 NAME                                                             READY   STATUS   AGE
package-operator-system   ServiceAccount/package-operator                                  -                5m31s
                          ├── ClusterObjectSet/package-operator-56496479f6                 -                5m20s
                          │   └── ClusterObjectDeployment/package-operator                 -                5m21s
                          │       └── ClusterPackage/package-operator                      -                5m22s
                          ├── ClusterRoleBinding/package-operator                          -                5m31s
                          │   ├── ClusterObjectSet/package-operator-56496479f6             -                5m20s
                          │   └── ClusterRole/cluster-admin                                -                6m56s
                          ├── ClusterRoleBinding/system:basic-user                         -                6m56s
                          │   └── ClusterRole/system:basic-user                            -                6m56s
                          ├── ClusterRoleBinding/system:discovery                          -                6m56s
                          │   └── ClusterRole/system:discovery                             -                6m56s
                          ├── ClusterRoleBinding/system:public-info-viewer                 -                6m56s
                          │   └── ClusterRole/system:public-info-viewer                    -                6m56s
                          └── ClusterRoleBinding/system:service-account-issuer-discovery   -                6m56s
                              └── ClusterRole/system:service-account-issuer-discovery      -                6m56s

and also a deployment kubectl lineage -D -n package-operator-system Deployment/package-operator-manager:

NAMESPACE                 NAME                                               READY   STATUS   AGE
package-operator-system   Deployment/package-operator-manager                1/1              11m
                          └── ClusterObjectSet/package-operator-56496479f6   -                11m
                              └── ClusterObjectDeployment/package-operator   -                11m
                                  └── ClusterPackage/package-operator        -                11m

I'll link the Package sources for reference [2]. The full list of missing dependent objects should be:

Deployment/package-operator-manager
ServiceAccount/package-operator

All objects have proper ownerReferences set up. I've tried to include all relevant object manifest in the attached dump. [3]

[1]

kubectl apply -f https://github.com/package-operator/package-operator/releases/download/v1.6.5/self-bootstrap-job.yaml

[2] https://github.com/package-operator/package-operator/tree/v1.6.6/config/packages/package-operator
[3] package-operator-objectdump.tar.gz

A lot of cves

error: unable to retrieve the complete list of server APIs: metrics.k8s.io/v1beta1: Unauthorized

Similar issue fixed in helm: helm/helm#6361

Allow excluding events

i love the plugin and use it alot! the one issue i have is that when running this on a long living object which either it or dependent objects have lots of historical events, the output is unclear and an event can and very often is tansient (for example waiting for PV etc.).

is it possible to have a flag that would allow filtering out resources shown in the tree something like:
kubectl lineage deployment/demo-app --exclude=events

or even one that would be specific to events:
kubectl lineage deployment/demo-app --exclude-events=true