I use Laravel Horizon in Kubernetes with HPA, which allows me to have several worker Pods.
But during a scalldown or rollout update, all workers in all pods are restarted.
I had to create my own terminate command to remove the queue:restart cache insertion but I was wondering if it really had a purpose or if it could be removed directly from the package.
Steps To Reproduce
Start several Horizon using the same redis and terminate one of them. Workers in other Horizon will also quit and restart.
Horizon Version
5.21.1
Laravel Version
10.4
PHP Version
8.2
Redis Driver
PhpRedis
Redis Version
5
Database Driver & Version
No response
Description
I don't understand why a full worker restart is requested when using the
horizon:terminate
command.You can see this line Here
I use Laravel Horizon in Kubernetes with HPA, which allows me to have several worker Pods. But during a scalldown or rollout update, all workers in all pods are restarted.
I had to create my own terminate command to remove the queue:restart cache insertion but I was wondering if it really had a purpose or if it could be removed directly from the package.
Steps To Reproduce
Start several Horizon using the same redis and terminate one of them. Workers in other Horizon will also quit and restart.