WatchTower: The Datadog Agent Alternative That Just Works

The Problem with datadog/agent

Datadog’s Docker agent is one of the most popular monitoring images on Docker Hub with 1 billion+ pulls. But it has fundamental problems:

1. Vendor Lock-in

You cannot use the Datadog agent without an API key. Your data goes to Datadog’s cloud. Period. There’s no self-hosted mode.

2. No Built-in Dashboard

The agent only collects metrics. You still need Grafana (or Datadog’s expensive cloud dashboard) to visualize them.

3. Complex Configuration

docker run -d --name datadog-agent \
  -e DD_API_KEY=your-key \
  -e DD_SITE=datadoghq.com \
  -e DD_LOGS_ENABLED=true \
  -e DD_LOGS_CONFIG_CONTAINER_COLLECT_ALL=true \
  -e DD_APM_ENABLED=true \
  -e DD_PROCESS_AGENT_ENABLED=true \
  -e DD_CONTAINER_EXCLUDE="name:datadog-agent" \
  -v /var/run/docker.sock:/var/run/docker.sock:ro \
  -v /proc:/host/proc:ro \
  -v /sys/fs/cgroup:/host/sys/fs/cgroup:ro \
  datadog/agent:latest

That’s 7 environment variables just to get basic monitoring working.

4. Heavy Resource Usage

The Datadog agent image is 600MB+ and uses 200-400MB of RAM per host.

5. Known Bugs

• Container collection forced on even when disabled
• Architecture mismatches (linux/amd64 missing from tags)
• Docker label configuration bugs
• Log port specification issues

Enter WatchTower

WatchTower is a drop-in replacement that fixes all of these problems.

One Command Setup

docker run -d \
  --name watchtower \
  -v /var/run/docker.sock:/var/run/docker.sock \
  -p 3000:3000 -p 9090:9090 \
  ghcr.io/irfancode/watchtower:latest

That’s it. Zero environment variables. Open http://localhost:3000 and you have a live dashboard.

What You Get

Built-in Dashboard:

• Real-time container CPU, memory, network usage
• Container status (running/stopped)
• Restart counts
• Auto-refresh every 15 seconds

Prometheus Metrics:

• container_cpu_percent — Per-container CPU usage
• container_memory_mb — Per-container memory usage
• container_network_rx_bytes / container_network_tx_bytes — Network I/O
• container_status — Running (1) or stopped (0)
• total_containers / running_containers — Cluster overview

Comparison

When to Use WatchTower

• You want container monitoring without vendor lock-in
• You need a quick dashboard without setting up Grafana
• You’re running Docker on a single host or small cluster
• You want Prometheus-compatible metrics
• You’re cost-conscious

When to Stick with Datadog

• You need APM (distributed tracing)
• You need log management at massive scale
• You have a dedicated budget and want vendor SLAs
• You need compliance certifications (SOC2, HIPAA)

Getting Started

# Basic usage
docker run -d --name watchtower \
  -v /var/run/docker.sock:/var/run/docker.sock \
  -p 3000:3000 -p 9090:9090 \
  ghcr.io/irfancode/watchtower:latest

# With custom settings
docker run -d --name watchtower \
  -v /var/run/docker.sock:/var/run/docker.sock \
  -p 3000:3000 -p 9090:9090 \
  -e SCRAPE_INTERVAL=10s \
  -e CONTAINER_FILTER=running \
  ghcr.io/irfancode/watchtower:latest

Docker Compose

services:
  watchtower:
    image: ghcr.io/irfancode/watchtower:latest
    container_name: watchtower
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock:ro
    ports:
      - "3000:3000"
      - "9090:9090"
    environment:
      - SCRAPE_INTERVAL=10s
    restart: unless-stopped

The Bottom Line

WatchTower isn’t trying to replace every feature of Datadog. It’s trying to replace the 80% of teams who just need to know: “Are my containers healthy, and how much resources are they using?”

For that use case, it’s simpler, cheaper, and more private than the alternative.

GitHub: github.com/irfancode/docker-toolkit/watchtower License: MIT