containerssh / libcontainerssh Goto Github PK

Describe the bug

The tests for the sshproxy module are flaky and sometimes meet their end at the hands of the 10m timeout.

To Reproduce

Not reproducible.

Expected behavior

The sshproxy tests should be stable.

Please describe what you would like to see in ContainerSSH

The remote server IP in ssh-proxy backend may change with the config webhook. We cannot get the current connections for a specific server. Thus I propose to add a new label server-ip to the metric containerssh_ssh_current_connections

Please describe your use case

I will create multiple ssh backends. And use them in ssh-proxy mode. The server IP is updated by the config server webhook. And I wanna know how many connections are alive in the server.

Please describe what you would like to see in ContainerSSH

Currently ContainerSSH provides a webhook for authentication and configuration of a container (or pod) that are triggered by an SSH connection to the ContainerSSH server. Upon successful authentication, the container described by the configuration is instantiated. Upon SSH disconnect, the container is removed by the server.

We would like a third webhook which would receive the status of the removal.

Please describe your use case

In order to run a ConatinerSSH created container successfully in a k8s environment, we choose to spin up auxiliary services (VMs, pods, etc). We would like some signal to know when we can tear down those extra resources.

Please describe what you would like to see in ContainerSSH

sshproxy:
  server: localhost
  port: 22222
  username: envd
  privateKey: /home/gaocegege/.config/envd/id_rsa_envd
  allowedHostKeyFingerprints:
   - SHA256:

The allowedHostKeyFingerprints cannot be empty now. It will be better to add a new config to disable the host key fingerprints check

Please describe your use case

A clear and concise description of how you would use this feature.

What would you like to add/change in the documentation?

Investigate the behaviour of struct merging, especially when it pertains to lists, when returning results from the configuration server and the configs have to be merged. A user reported that lists with multiple items overrode entire objects if e.g. only the 3rd item needed updating.

The underlying backend is the mergo library.

Where did you try to find the information?

https://containerssh.io/reference/configserver/?h=server

Describe the bug

go mod init container-ssh/config-server
go get github.com/containerssh/libcontainerssh
go mod tidy

fails with

container-ssh/config-server imports
	github.com/containerssh/kubernetes imports
	k8s.io/client-go/kubernetes/scheme imports
	k8s.io/api/batch/v2alpha1: module k8s.io/api@latest found (v0.29.2), but does not contain package k8s.io/api/batch/v2alpha1
container-ssh/config-server imports
	github.com/containerssh/kubernetes imports
	k8s.io/client-go/kubernetes/scheme imports
	k8s.io/api/discovery/v1alpha1: module k8s.io/api@latest found (v0.29.2), but does not contain package k8s.io/api/discovery/v1alpha1
container-ssh/config-server imports
	github.com/containerssh/kubernetes imports
	k8s.io/client-go/kubernetes/scheme imports
	k8s.io/api/flowcontrol/v1alpha1: module k8s.io/api@latest found (v0.29.2), but does not contain package k8s.io/api/flowcontrol/v1alpha1
container-ssh/config-server imports
	github.com/containerssh/kubernetes imports
	k8s.io/client-go/rest imports
	k8s.io/apimachinery/pkg/util/clock: module k8s.io/apimachinery@latest found (v0.29.2), but does not contain package k8s.io/apimachinery/pkg/util/clock

To Reproduce

see above

Expected behavior

Dependencies are successfully resolved

Version

Try to build a config-server for containerssh 0.5 with go 1.22.0 (worked fine under go 1.17 and containerssh 0.4)

Workaround

I'm not sure if I missed something. When I add a

require (
     k8s.io/client-go v0.29.2
)

before calling "go mod tidy" there is no error logged.

Change the KinD utility download to:

1. Perform a hash check
2. Don't run curl with sudo, move the kind utility later

Please describe what you would like to see in ContainerSSH

Support multiple audit log storage in one containerssh instance.

Please describe your use case

I expect to keep the audit log in s3 and the local file system.

Describe the bug

Prometheus seems to require all metrics to end with the unit name.

Error: unit not a suffix of metric "containerssh_config_server_failures"

To Reproduce

If you can please provide the steps to reproduce the issue.

1. Run ContainerSSH with metrics enabled
2. Setup a kubernetes ServiceMonitor to scrape ContainerSSH services

Expected behavior

ContainerSSH metrics are imported into Prometheus.

Version

0.5

Please describe what you would like to see in ContainerSSH

go lang version up 1.18 -> 1.21

some package depandency upgrade

Please describe your use case

A clear and concise description of how you would use this feature.