nsq-operator targets to run NSQ-As-A-Service, i.e., NAAS, based on Kubernetes.
nsq-operator is currently in Alpha stage. Until GA status is clearly stated, it is not recommended to run nsq-operator in production.
But, we encourage end users to give nsq-operator a try, i.e., testing and benching nsq-operator, and give us some feedback.
- High availability. NSQ-Operator supports HA by utilizing Kubernetes leaderelection package and coordination.k8s.io/lease resource.
- Create/Delete a nsq cluster
- Scale nsqd/nsqlookupd/nsqadmin components separately
- Update nsqd/nsqlookupd/nsqadmin image separately
- Adjust nsqd command line arguments
- Adjust nsqd pods memory resource according to average message size, memory queue size, memory overbooking percent and channel count
- Log Management.
- Rotate log by logrotate hourly
- Mount log directory to a dedicated host machine directory in nsqd/nsqlookupd/nsqadmin spec
- QPS based horizontal pod autoscale. No memory/cpu/network/disk resources based autoscale support. QPS autoscale does not rely on any other systems. It is implemented in Kubernetes way. More details
- Kubernetes >= 1.14.0
- Golang >= 1.12.0
There is a sdk which can be used to CRUD nsq related resource. Files under examples show sdk use examples.
| File | Description |
|---|---|
| create_cluster.go | Create a NSQ cluster |
| delete_cluster.go | Delete a NSQ cluster |
| update_nsqadmin_replica.go | Update nsqadmin resource object replica |
| update_nsqlookupd_replica.go | Update nsqlookupd resource object replica |
| update_nsqd_replica.go | Update nsqd resource object replica |
| adjust_nsqd_config.go | Adjust nsqd command line arguments |
| adjust_nsqd_memory_resource.go | Adjust nsqd resource object's memory resource limits/requests |
| adjust_nsqdscale.go | Adjust nsqdscale resource object |
| bump_nsqadmin_image.go | Update a NSQ cluster's nsqadmin image |
| bump_nsqlookupd_image.go | Update a NSQ cluster's nsqlookupd image |
| bump_nsqd_image.go | Update a NSQ cluster's nsqd image |
See FAQ
See Users