PowerScale Data Insights

Collect, store, and visualize Dell PowerScale OneFS cluster performance data.

PowerScale Data Insights is the successor to the Isilon Data Insights Connector. It replaces the Python-based collector with two purpose-built Go collectors, adds a dashboard generator for Partitioned Performance datasets, and ships modernized Grafana dashboards.

Components

Component	Description
gostats	Collects OneFS statistics via PAPI and writes to InfluxDB v1/v2 or Prometheus
goppstats	Collects OneFS Partitioned Performance data via PAPI and writes to InfluxDB v1/v2 or Prometheus
dashgen	Generates Grafana dashboards for Partitioned Performance datasets by querying the PAPI at runtime

Architecture

┌──────────────────┐     ┌──────────────────┐
│  OneFS Cluster 1 │     │  OneFS Cluster N │
│       (PAPI)     │     │       (PAPI)     │
└────────┬─────────┘     └────────┬─────────┘
         │                        │
    ┌────┴────────────────────────┴────┐
    │                                  │
    ▼                                  ▼
┌──────────┐                    ┌────────────┐
│ gostats  │                    │ goppstats  │
│(stats)   │                    │(PP data)   │
└────┬─────┘                    └─────┬──────┘
     │                                │
     └───────────┬────────────────────┘
                 │
                 ▼
          ┌────────────────┐
          │   InfluxDB     │
          │  (v1 or v2)    │
          │       or       │
          │  Prometheus    │
          └───────┬────────┘
                  │
                  ▼
          ┌────────────────┐
          │    Grafana     │
          │ (dashboards)   │
          └────────────────┘

Both collectors run independently, querying one or more OneFS clusters via PAPI and writing time-series data to the configured backend. Grafana reads from the time-series database and renders the pre-built dashboards. See docs/architecture.md for details.

Requirements

• Go 1.25+ (build from source)
• OneFS 9.x+ (PAPI v10 for Partitioned Performance, PAPI v3 for summary stats)
• InfluxDB v1.8+ or v2.x (InfluxQL compatibility)
• Grafana 10+ (legacy dashboard format with modern panel types)

Quick Start

There are three ways to get up and running:

Option 1: Docker Compose (fastest)

Brings up InfluxDB, Grafana, and both collectors in one command.

cd docker/
cp gostats.example.toml gostats.toml
cp goppstats.example.toml goppstats.toml
# Edit both files: set hostname, username, password under [[cluster]]

docker compose up -d

Open Grafana at http://localhost:3000 (admin/admin). Pre-built dashboards are provisioned automatically.

Option 2: Build from source

make build

This produces three binaries in bin/:

• bin/gostats
• bin/goppstats
• bin/dashgen

Configure and run:

cp configs/gostats.example.toml idic.toml
cp configs/goppstats.example.toml goppstats.toml
# Edit both files with your cluster and InfluxDB details

./bin/gostats -config-file idic.toml
./bin/goppstats -config-file goppstats.toml

Option 3: Docker (standalone containers)

# Build
docker build -f docker/Dockerfile.gostats -t pdi-gostats .
docker build -f docker/Dockerfile.goppstats -t pdi-goppstats .

# Run (mount your config file)
docker run -d -v /path/to/idic.toml:/etc/gostats/idic.toml pdi-gostats
docker run -d -v /path/to/goppstats.toml:/etc/goppstats/goppstats.toml pdi-goppstats

Generate a Partitioned Performance dashboard

./bin/dashgen -host <cluster> -user <user> -password <pass> \
  -dataset <id> -out pp-dashboard.json

Import the generated JSON into Grafana, or use the pre-built dashboards in dashboards/influxdb/.

Documentation

• Architecture — component design, data flow, shared library
• Getting Started — end-to-end setup guide
• Configuration Reference — complete TOML reference
• Dashboards — dashboard guide and using dashgen
• Deployment — binary, Docker, Docker Compose, Kubernetes
• OneFS Setup — cluster user creation and privileges
• Migrating from v1 — guide for Python connector users

Project Structure

powerscale_data_insights/
├── cmd/
│   ├── gostats/          Statistics collector
│   ├── goppstats/        Partitioned Performance collector
│   └── dashgen/          Dashboard generator
├── internal/             Shared library (PAPI client, backends, config, logging)
├── dashboards/influxdb/  Pre-built Grafana dashboards (InfluxQL)
├── configs/              Example configuration files
├── docker/               Dockerfiles, Docker Compose, Grafana provisioning
├── docs/                 Documentation
└── papi/                 OneFS PAPI schema reference

License

MIT