Operating DPS#

Day-to-day procedures for a running DPS deployment, grouped by what you are trying to do. If a command fails, refer to Troubleshooting.

For an end-to-end scenario rather than an individual operation, refer to MaxLPS Power Pilot Runbooks.

Change what DPS manages#

The operations you perform most often: adjusting which hardware DPS knows about, and which workloads hold power.

Managing Resource Groups — create and manage the groups that hold power for a workload
Resource Group Priority — set --priority and understand steal versus giveback order
Managing Devices — manage the devices in your deployment
Import a Topology — validate and import a topology from JSON through the CLI or API
Managing Topologies — work with topology configurations
Add nodes to DPS — bring new hardware into an existing deployment
DPS UI Operator Guide — create, validate, and inspect these workflows in the DPS UI

Telemetry and Observability — the metrics DPS exposes and the queries that make them useful
Provisioned and Observed Power — how to read the gap between committed and actual power, and what a persistent gap tells you
Prometheus monitoring — collect and visualize power metrics

Run a BMC Health Check — run read-only or approved write-enabled BMC health checks and interpret the report
Troubleshooting — operator diagnosis steps and escalation criteria

Less frequent, but the operations where being wrong is expensive.

Backing up with Velero — back up DPS with Velero
Credentials and Secrets Configuration — configure authentication and secrets management
Database TLS — configure PostgreSQL TLS, including CA verify and mutual TLS client certificates

Covered in full by Integrating DPS. The procedures themselves: