PANOSETI Grafana Provisioning

This repository manages the "Infrastructure as Code" (IaC) configuration for PANOSETI's Grafana instance. It ensures that our dashboards and datasources are version-controlled, reproducible, and recoverable.

Workflow Overview

Grafana does not support "Two-Way Sync" natively. We use a "ClickOps to Git" workflow:

1. Provision: Grafana launches and loads dashboards from the dashboards/ directory.
2. Edit: Users edit dashboards in the Grafana UI (enabled via allowUiUpdates: true).
3. Sync: An administrator runs sync_grafana.py to pull UI changes back to the local dashboards/ JSON files.
4. Commit: Changes are committed to Git.

Setup

1. Install Dependencies

The sync script requires Python 3 and a few libraries:

pip install requests python-slugify

2. Create a Service Account

To allow the script to talk to Grafana, you need a Service Account Token.

1. Log in to Grafana (Admin).
2. Go to Administration > Users and access > Service Accounts.
3. Click Add service account.

• Name: SyncBot
• Role: Viewer (Safe for pulling) or Admin (If pushing is needed later).

4. Click Add token -> Generate token.
5. Copy the token (e.g., glsa_...).

Usage

The Sync Script (sync_grafana.py)

This script compares the live Grafana dashboards with your local JSON files.

Basic Usage (Dry Run): By default, the script will not write to disk. It only shows you what changed.

export GRAFANA_API_KEY="glsa_YOUR_TOKEN_HERE"
python3 sync_grafana.py

Viewing Diffs: To see exactly which lines changed in the JSON:

python3 sync_grafana.py --diff

Saving Changes (The "Sync"): To actually overwrite your local files with the live versions:

python3 sync_grafana.py --sync

CLI Options

Flag	Description
--key <token>	Service Account Token (Can also use GRAFANA_API_KEY env var)
--url <url>	Grafana URL (Default: http://localhost:3000)
--dir <path>	Output directory (Default: ./grafana_provisioning/dashboards)
--sync	Required to save. If omitted, runs in Read-Only mode.
--diff	Prints colored line-by-line diffs of changes.

Directory Structure

• datasources/: YAML configurations for InfluxDB, Loki, etc.
• dashboards/: JSON models of the dashboards.
• sync_grafana.py: The bridge tool.
• Dockerfile: Builds the custom Grafana image with these files pre-loaded.