Skip to main content

Deployment

The Ops Platform supports two deployment forms: Docker Compose standalone deployment and Kubernetes cluster deployment. Choose based on the deployment form of the HAP/HDP private environment being served. In most cases, keep the same deployment form to simplify network connectivity, log ingestion, and Kubernetes metrics collection.

Served HAP/HDP FormOps Platform Deployment FormDescription
HAP/HDP in standalone mode (Docker Compose)Ops Platform standalone deploymentThe Ops Platform runs as ops-allinone on a single server and is mainly used for host, middleware, container log, and service log monitoring
HAP/HDP in cluster mode (Kubernetes)Ops Platform cluster deploymentThe Ops Platform is deployed inside the HAP/HDP Kubernetes cluster and can directly collect metrics for nodes, Pods, Deployments, and other cluster objects
Kubernetes metrics require the Ops Platform to be deployed inside the monitored cluster

Standalone Ops Platform deployment can monitor hosts, middleware, container logs, and service logs, but does not directly collect Kubernetes node/Pod/Deployment metrics. To collect a HAP/HDP Kubernetes cluster, prioritize deploying the Ops Platform into that cluster. If one Ops Platform needs to manage multiple Kubernetes clusters, use collection component mode.

Resource requirements: 4C/8G or above. Plan disk capacity by deployment form: standalone deployment recommends 100G or above; the cluster PVC plan requests 130Gi by default (Prometheus 100 + ops-mongo 20 + Grafana 10, adjustable in 30-stateful.yaml). Alert configuration is stored in the built-in ops-mongo, so no extra database is required.

Account permissions for monitored middleware: Each monitored target must provide a read-only account. See Data Sources for detailed permissions.

Next Steps

After deployment and access verification, continue to connect monitored targets:

  1. Data Sources: Register hosts and middleware
  2. Node Exporter: Monitor servers outside the cluster
  3. Kubernetes Cluster: Collect node/container/cluster object status
  4. Service Logs: Let HAP/HDP microservices push logs to the Ops Platform
  5. Tracing: Let business services report traces

The Object Storage and Environment Variables pages under Reference are supplemental materials for on-demand lookup.