thank you!

The helm chart do not seems to handle horizontal scaling.
Following the https://docs.soketi.app/advanced-usage/horizontal-scaling/clustering documentation, I used the following env vars.

env:
- name: PORT
  value: "6001"
- name: MODE
  value: full
- name: CLUSTER_PORT
  value: "11002"
- name: SOKETI_DEBUG
  value: "1"
- name: ADAPTER_DRIVER
  value: cluster

I ended with 3 nodes that were elected master with the following logs:

soketi-66bfb59674-wmnxg soketi [Fri Mar 04 2022 22:44:16 GMT+0000 (Coordinated Universal Time)] Promoted from node to master.
soketi-66bfb59674-wmnxg soketi {
soketi-66bfb59674-wmnxg soketi   isMaster: true,
soketi-66bfb59674-wmnxg soketi   isMasterEligible: true,
soketi-66bfb59674-wmnxg soketi   weight: -0.1646433855526,
soketi-66bfb59674-wmnxg soketi   address: '127.0.0.1',
soketi-66bfb59674-wmnxg soketi   advertisement: undefined
soketi-66bfb59674-wmnxg soketi }

Is it normal or a configuration issue on my side or a known issue? Message are not always broadcaster to client, depending on which node is connected the client.

Hello,

Thank you for the effort of putting together a Helm chart, I was wondering if there's any example to expose the soketi service using Ingress controller?

What I've tried so far:

# https://kubernetes.io/docs/concepts/services-networking/ingress/#the-ingress-resource

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: ingress-nginx-ws
  namespace: {{ .Values.application.namespace }}
  annotations:
    nginx.ingress.kubernetes.io/rewrite-target: /$1
    nginx.ingress.kubernetes.io/backend-protocol: "FCGI"
    nginx.ingress.kubernetes.io/fastcgi-index: "index.php"
    nginx.ingress.kubernetes.io/fastcgi-params-configmap: "ingress-conf"
    nginx.ingress.kubernetes.io/websocket-services: php-soketi
    nginx.org/websocket-services: php-soketi
    nginx.ingress.kubernetes.io/proxy-read-timeout: "3600"
    nginx.ingress.kubernetes.io/proxy-send-timeout: "3600"

spec:
  ingressClassName: "nginx"
  rules:
  - http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: {{ .Values.service.name }}
            port:
              name: {{ .Values.service.port.name }}
      - path: /ws
        pathType: Prefix
        backend:
          service:
            name: php-soketi
            port:
              number: 6001

but I recieve the following error using Laravel Echo:

(index):29 WebSocket connection to 'ws://192.168.58.2/ws:80/app/cuG014DSf9tjcjyvvuOHNfhBSZqqOBsp?protocol=7&client=js&version=8.4.0-rc2&flash=false' failed: WebSocket is closed before the connection is established.

See here: https://docs.bitnami.com/kubernetes/apps/aspnet-core/administration/deploy-extra-resources/

I want to add a custom config file, but with the existing helm chart it's not possible to pass a custom config map.. Could you add a extraDeploy key?

Thank you! ❤️

Checked the https://helm.soketi.app/index.yaml to look for the 0.3.0 version and is not there.

Can you push it up?

Have the server running pretty good on local trying to have it running on my k8s :)

Thanks,
Francisco

What is this package usage ?

Can i show metrics of websocket transactions in my website ? (ex: Transfered messages per hour, connections per hour, etc)

Also is there any documentation to better understanding this package ?

Hello all,

I'm experiencing a strange behavior with this helm chart. For the truth this is the first time that something goes wrong with it, because in the past i've already used it without problems.

node:internal/validators:217 │ │ throw new ERR_SOCKET_BAD_PORT(name, port, allowZero); │ │ ^ │ │ │ │ RangeError [ERR_SOCKET_BAD_PORT]: Port should be > 0 and < 65536. Received udp://10.43.244.58:11002. │ │ at new NodeError (node:internal/errors:371:5) │ │ at validatePort (node:internal/validators:217:11) │ │ at Socket.send (node:dgram:637:12) │ │ at /app/node_modules/node-discover/lib/network.js:188:25 │ │ at Array.forEach (<anonymous>) │ │ at /app/node_modules/node-discover/lib/network.js:187:26 │ │ at Network.encode (/app/node_modules/node-discover/lib/network.js:221:12) │ │ at Network.send (/app/node_modules/node-discover/lib/network.js:180:10) │ │ at Discover.hello (/app/node_modules/node-discover/lib/discover.js:361:17) │ │ at /app/node_modules/node-discover/lib/discover.js:286:10 { │ │ code: 'ERR_SOCKET_BAD_PORT'

It seems that even with a values.yaml ( and without ) something get wrong with PORT.

Is this a bug? Or am I missing something?

Note : I'm using this in develop env through a k3d env, traefik based

thanks!

Hello,

When we launch the Socketi with Network watcher enabled, we get the below error:

Starting the watcher...
Namespace: dev
Pod name: soketi-soketii-5f845d945b-zj5mn
Server port: 6001
Memory threshold: 85%
Monitoring interval: 1s

In ChecksCurrentPod.php line 97:
                                                                               
  App\Commands\WatchNetworkCommand::getUsage(): Return value must be of type   
  array, null returned

After investigation, the following file / Line https://github.com/soketi/charts/blob/master/charts/soketi/templates/deployment.yaml#L76 is wrong and should not read the SERVER_PORT from .Values.service.port but instead should be the metrics server port 9601.

My suggestion is to add in values the following .Values.metrics.port and use it across the board where needed and in SERVER_PORT env of the network watcher.

I hope this helps everyone else as well!

I have installed k8 soketi via helm chart and enabled metrics and serviceMonitor
However, when I try to access my-k8-soketi-k8soketi-metrics service using url http://<IP>:9601/metrics it is not accessible via curl.

The metrics service in GKE is not having any endpoint.

EKS Cluster Version: 1.22
EKS Node Version: v1.22.17-eks-49d8fe8
Helm Version: version.BuildInfo{Version:"v3.11.2", GitCommit:"912ebc1cd10d38d340f048efaf0abda047c3468e", GitTreeState:"dirty", GoVersion:"go1.20.2"}

Steps to reproduce:

1. Deploying application using commands from READ.me
   
2. Getting success message.
   
3. Expecting working app.

But currently I have broken POD with the message below

node:internal/validators:226
    throw new ERR_SOCKET_BAD_PORT(name, port, allowZero);
    ^

RangeError [ERR_SOCKET_BAD_PORT]: Port should be > 0 and < 65536. Received udp://172.20.63.90:11002.
    at new NodeError (node:internal/errors:387:5)
    at validatePort (node:internal/validators:226:11)
    at Socket.send (node:dgram:620:12)
    at /app/node_modules/node-discover/lib/network.js:188:25
    at Array.forEach (<anonymous>)
    at /app/node_modules/node-discover/lib/network.js:187:26
    at Network.encode (/app/node_modules/node-discover/lib/network.js:221:12)
    at Network.send (/app/node_modules/node-discover/lib/network.js:180:10)
    at Discover.hello (/app/node_modules/node-discover/lib/discover.js:361:17)
    at /app/node_modules/node-discover/lib/discover.js:286:10 {
  code: 'ERR_SOCKET_BAD_PORT'
}

Current POD status is
Restarting failing container

Getting an error when deploying this chart to gke:

2022-02-15 14:21:35.906 CSTnode:internal/validators:216
Error
2022-02-15 14:21:35.906 CST throw new ERR_SOCKET_BAD_PORT(name, port, allowZero);
Error
2022-02-15 14:21:35.906 CST ^
Error
2022-02-15 14:21:35.906 CST{}
Error
2022-02-15 14:21:35.906 CSTRangeError [ERR_SOCKET_BAD_PORT]: Port should be > 0 and < 65536. Received udp://10.124.4.184:11002.
Error
2022-02-15 14:21:35.907 CST at new NodeError (node:internal/errors:371:5)
Error
2022-02-15 14:21:35.907 CST at validatePort (node:internal/validators:216:11)
Error
2022-02-15 14:21:35.907 CST at Socket.send (node:dgram:637:12)
Error
2022-02-15 14:21:35.907 CST at /app/node_modules/node-discover/lib/network.js:188:25
Error
2022-02-15 14:21:35.907 CST at Array.forEach (<anonymous>)
Error
2022-02-15 14:21:35.907 CST at /app/node_modules/node-discover/lib/network.js:187:26
Error
2022-02-15 14:21:35.907 CST at Network.encode (/app/node_modules/node-discover/lib/network.js:221:12)
Error
2022-02-15 14:21:35.907 CST at Network.send (/app/node_modules/node-discover/lib/network.js:180:10)
Error
2022-02-15 14:21:35.907 CST at Discover.hello (/app/node_modules/node-discover/lib/discover.js:361:17)
Error
2022-02-15 14:21:35.907 CST at /app/node_modules/node-discover/lib/discover.js:286:10 {
Error
2022-02-15 14:21:35.907 CST code: 'ERR_SOCKET_BAD_PORT'

Hey

This is to add topologySpreadConstraints support for helm chart

https://kubernetes.io/docs/concepts/scheduling-eviction/topology-spread-constraints/

Sorry for probably missing something really obvious - but how do you do the setup the values?

I can see the avilable variables here: https://docs.soketi.app/app-management/array-driver

But when installing with helm, how do I pass SOKETI_DEFAULT_APP_ID, SOKETI_DEFAULT_APP_KEY, SOKETI_DEFAULT_APP_SECRET? Thanks

TIA