Introduction

Proxima Console

AI-powered infrastructure operations — inventory, monitoring, compliance, terminal access, and AI-assisted ops, all from one console.

Get Started →

Architecture

API Reference

---

Get Started

1. Install the Agent

Deploy the lightweight Go agent on your servers. It auto-discovers inventory, collects metrics, and connects via NATS.

Agent Installation →

2. Connect & Monitor

View hosts in the dashboard with live CPU, memory, and disk metrics. Set up healthchecks and compliance rules.

Monitoring →

3. Operate with AI

Use the CLI (pc ssh), run runbooks, and ask Claude about your infrastructure via AI Chat or MCP.

Terminal CLI → · AI Chat →

---

Platform Features

Inventory & Discovery

• Host inventory — OS, packages, services, containers, hardware
• Kubernetes clusters — 13 resource types
• Database topology — PostgreSQL primary/replica detection
• Asset management (CMDB)

Monitoring & Compliance

• System metrics — CPU, memory, disk, network, processes
• Integration collectors — PostgreSQL, Nginx, Docker
• Compliance scoring — healthcheck rules, frameworks
• Alert correlation — Alertmanager ingestion + enrichment

Terminal Access & CLI

• Browser terminal — WebSocket with xterm.js
• pc CLI tool — SSH, SCP, port forwarding, session join
• Session recording — asciicast v2 with playback
• Multi-party sessions — observe or participate

AI & Automation

• AI Chat — Claude-powered with 20+ infrastructure tools
• MCP Server — for Claude Desktop, Cursor, VS Code
• Runbooks — YAML DSL with 10 typed actions
• Service Desk — Jira ITSM integration

Security & Access

• RBAC — roles, permissions, client scoping
• Passkeys (WebAuthn) — passwordless login
• TOTP MFA — two-factor authentication
• Audit logging — all API operations tracked

Observability

• Centralized logs — VictoriaLogs + LogsQL
• Distributed tracing — OpenTelemetry + Tempo
• Grafana dashboards — auto-provisioned
• Change detection — file integrity + config drift

---

Architecture

Managed Hosts              Proxima Platform                    Clients
┌──────────┐          ┌─────────────────────┐          ┌──────────────┐
│  Agent   │──NATS──▶ │  Backend (Go API)   │ ◀─REST── │  Frontend    │
│  (Go)    │          │  PostgreSQL          │          │  (React/TS)  │
│          │          │  VictoriaMetrics     │          │              │
└──────────┘          │  NATS + JetStream    │          └──────────────┘
                      │  Vault · Valkey      │          ┌──────────────┐
┌──────────┐          │                     │ ◀─gRPC── │  pc CLI      │
│  Agent   │──NATS──▶ │  Tempo · VLogs      │          │  (Go)        │
│  (K8s)   │          └─────────────────────┘          └──────────────┘
└──────────┘                    ▲                       ┌──────────────┐
                                │ REST                  │  MCP Server  │
                                └───────────────────────│  (Go)        │
                                                        └──────────────┘

• Agent — lightweight Go binary on every managed server. Collects inventory, metrics, runs healthchecks.
• Backend — REST API with chi router, RBAC, workers for metric/log processing.
• Frontend — React SPA with Tailwind + shadcn/ui.
• pc CLI — native terminal access via gRPC, host listing, file transfer, port forwarding.
• MCP Server — exposes infrastructure data as tools for LLM integration.

Architecture deep dive → · Data flow →

---

What's New in v0.6.x

• L1 incident agent — escalation engine, on-call schedules, deterministic triage context packs, AI-assisted RCA, a human-in-the-loop memory loop, and a Telegram ops bot
• Agent fleet management — version-skew visibility and canary/batched self-update rollouts with pause/resume/abort and automatic pause-on-failure
• MCP expansion — action tools, structured output, browseable resources, and invokable prompts (triage/diagnose/compliance/fleet workflows)
• Agent distributed tracing & self-log shipping — OTel spans and WARN+ logs routed over NATS to the backend (UI-controlled)
• Observability alerting — Grafana Alerting fires high-signal rules routed to Telegram
• Daily briefing home page — a prioritized "Needs attention" view on the console root
• Per-client feature flags — gate features per client with launch defaults and an admin toggle UI

Full changelog → · Product roadmap →

---

Quick Links

I want to...	Go here
Install the agent on a server	Agent Installation
Install the pc CLI tool	CLI Reference
Set up monitoring	Monitoring Overview
Connect a K8s cluster	Cluster Agent
Configure RBAC roles	Roles & Permissions
Use AI Chat	AI Chat Overview
Set up MCP for Claude Desktop	MCP Setup
Create a runbook	Creating Runbooks
Troubleshoot issues	Troubleshooting Guide
View the API	API Endpoints