cramd/lookingglass repository overview

Looking Glass: Optimized Observability for Proxmox VE

A self-hosted, lightweight telemetry console designed specifically for Proxmox VE clusters.

Looking Glass serves as an optimized frontend for the ShiftMon stack (VictoriaMetrics + VictoriaLogs). It provides a unified pane of glass for your virtualized infrastructure, featuring:

⚡ Real-time instrumentation without the bloat.

📜 Lightning-fast log streaming directly from VictoriaLogs.

🤖 Contextual runbook diagnostics to instantly troubleshoot alerts.

🪶 A minimal resource footprint engineered to save your compute for actual workloads.

Architecture Overview

Standard monitoring patterns based on full-scale Prometheus and Grafana deployments often introduce significant overhead. Looking Glass addresses this by acting as a streamlined administrative console that interfaces directly with your TSDB backend.

Rather than utilizing persistent, high-frequency local scraping agents on the hypervisor, Looking Glass capitalizes on Proxmox VE's built-in metric servers:

Metrics: Native InfluxDB Push Metric Protocol -> VictoriaMetrics

Logs: Telegraf Syslog Integration (RFC 5424) -> VictoriaLogs

In baseline production testing, this architecture demonstrated up to a 92% reduction in memory footprint compared to standard Prometheus installations.

Key Technical Advantages

Resource Efficiency: Sub-second syslog pipeline latency and drastically reduced RAM overhead compared to traditional TSDB architectures.

Logical Environment Isolation: Group hypervisor nodes into custom environments (Production, Staging, Lab) to filter dashboards, logs, and alerts instantly using client-side caching.

Contextual Runbook Diagnostics: Cross-references active system alerts (via vmalert and Alertmanager) with local operational runbooks, providing actionable remediation steps directly in the console.

Platform Interoperability: Includes native deep links to pivot directly from aggregated metrics into raw Grafana dashboards, VictoriaMetrics VMUI, or the VictoriaLogs engine for advanced querying.

Quick Start (Docker Compose)

Deploy Looking Glass alongside your existing ShiftMon stack using Docker Compose.

Create a docker-compose.yml file:

`version: "3.8" services: backend: image: shiftsystems/looking-glass-backend:latest ports: - "5000:5000" env_file: - .env restart: unless-stopped

frontend: image: shiftsystems/looking-glass-frontend:latest ports: - "3000:3000" environment: - NEXT_PUBLIC_API_URL=http://localhost:5000⁠ restart: unless-stopped`

Create a .env file in the same directory to configure your target endpoints:

`# Target Infrastructure Endpoints VM_CLOUD_URL=http://192.168.0.100:8428⁠ VM_DIRECT_URL=http://192.168.0.100:8428⁠

⁠Core Telemetry Processors

VICTORIALOGS_URL=http://192.168.0.100:9428⁠ VMALERT_URL=http://192.168.0.100:8880⁠ ALERTMANAGER_URL=http://192.168.0.100:9093⁠

⁠System Execution Limits (free | pro | enterprise)

SAAS_USER_TIER=free PORT=5000`

Start the stack:

docker-compose up -d

The console will be accessible at http://localhost:3000⁠.

License & Project Links

License: Apache 2.0 / MIT

Documentation & Source: Looking Glass on GitHub

Parent Project: HoneOS