Roadmap

Where we're headed.

Overwatch is built in the open. Here's what's coming next on the hosted product: OpenTelemetry ingest, MCP for agents, and an on-prem collector for enterprise infrastructure metrics.

Need to run checks yourself today? The open-source core already covers self-hosted monitoring today. The roadmap below is for the hosted cloud product.

Planned

OpenTelemetry Ingest

Planned

Ship logs, metrics, and traces from your infrastructure or application directly to Overwatch using the OpenTelemetry standard. Correlate check results with application-level telemetry for full-stack visibility.

▸Native OTLP/gRPC and OTLP/HTTP endpoints
▸Logs, metrics, and traces in a unified timeline
▸Correlate check failures with application errors
▸Compatible with any OTEL SDK or collector
▸Configurable retention and sampling policies

  app ──▶ OTLP ──┐
                  │
  infra ──▶ OTLP ──┼──▶ ▲ OVERWATCH ──▶ correlate
                  │
  k8s ──▶ OTLP ──┘

MCP & AI-First Root-Cause Analysis

Planned

Hook your agent up to Overwatch over MCP so it can query checks, alert history, logs, metrics, and retention windows in natural language — then get a clear root-cause view when something breaks.

▸MCP server for agent access to monitors, history, and telemetry
▸Natural-language root-cause queries across signals
▸Automatic service dependency mapping
▸Correlate deployments, config changes, and incidents
▸Actionable remediation suggestions

  ┌──────────────────────────────┐
  │  "Why is checkout slow?"     │
  └──────────┬───────────────────┘
             ▼
  ┌──────────────────────────────┐
  │ ▲ OVERWATCH MCP              │
  │  deploy@14:02 ──▶ db p95 ↑   │
  │  svc-payment ──▶ timeout     │
  │  root: connection pool sat.  │
  └──────────────────────────────┘

On-Prem Collector

Planned

An optional collector that sits on your network, scrapes traditional infrastructure sources, and ships metrics into Overwatch — so VMware, Windows WMI, and other enterprise systems land in the same place as your uptime checks.

▸On-premises collector with outbound-only connectivity
▸Windows performance counters via WMI
▸VMware vSphere / ESXi host and VM metrics
▸CPU, memory, disk, network out of the box
▸Path for other scrapers via a plugin-style collector model

  ┌─────────┐  ┌─────────┐
  │ WINDOWS │  │ VMWARE  │
  │   WMI   │  │ vSphere │
  └────┬────┘  └────┬────┘
       └──────┬─────┘
              ▼
      ┌──────────────┐
      │  COLLECTOR   │ ── outbound ──▶ ▲ OVERWATCH
      └──────────────┘
         metrics · alerts · trends

  ┌───────────────────────────────────────┐
  │                                       │
  │   Ship with us. Build with us.        │
  │                                       │
  └───────────────────────────────────────┘

Ship with us. Build with us.

Want to shape what's next?

Start a free trial today, or tell us what matters most to your team. We prioritize based on real feedback.

Start free trial →Share feedback