Deployment
The Ops Platform supports Docker Compose standalone deployment and Kubernetes cluster deployment. Usually, match the HAP/HDP deployment method to simplify network connectivity, log ingestion, and Kubernetes metrics collection.
| HAP/HDP Deployment Method | Ops Platform Deployment Method | Description |
|---|---|---|
| Standalone (Docker Compose) | Standalone Deployment | Run ops-allinone on a single server to monitor hosts, middleware, container logs, and service logs |
| Cluster (Kubernetes) | Cluster Deployment | Deploy inside the HAP/HDP Kubernetes cluster to collect metrics for nodes, Pods, Deployments, and other cluster resources directly |
When deployed inside the monitored Kubernetes cluster, the Ops Platform can collect its metrics directly. When deployed outside the target cluster or monitoring multiple clusters, deploy the collection components inside each target cluster.
Docker Compose deployment requires at least 4 available CPU cores, 8 GB of memory, and 200 GB of disk space. The Kubernetes hostPath deployment node requires at least 200 GB of free disk space; the PVC deployment requests 250 GiB by default, adjustable in 30-stateful.yaml. Alert configuration is stored in the Ops Platform's own MongoDB, so no separate database is required.
Accounts for monitored middleware require appropriate permissions. Requirements depend on the data source type and purpose. See Data Sources.
Next Steps
After deployment and access verification, connect the monitored targets:
- Data Sources: Register hosts and middleware
- Node Exporter: Monitor servers outside the cluster
- Kubernetes Cluster: Collect node, container, and cluster resource status
- Service Logs: Configure HAP/HDP microservices to push logs to the Ops Platform
- Tracing: Configure business services to report traces
See Object Storage and Environment Variables under Reference for additional information as needed.