My open‑source monitoring stack in 2026 – 5 tools that replaced Datadog for $0

Description and use

The author swapped Datadog for a self‑hosted stack (Prometheus, Grafana, Loki, SigNoz, Uptime Kuma), cutting costs from ~$900 to ~$10/month, at the expense of managed reliability and instant alerts.

Problem

Datadog cost $900–$1,500/month for a 12‑service, 20‑host setup with 50 GB/day metrics, which was unsustainable for a side project.

Outcome

The stack reduced monthly cost from ~$900 to ~$10 while retaining metrics, logs, APM, uptime, and alerting, but required self‑hosted maintenance and manual instrumentation.

Workflow

• Deploy a single VM with 4 GB RAM
• Install Prometheus (≈1 GB RAM) and Grafana (≈200 MB)
• Configure Prometheus to scrape Node Exporter and app exporters
• Set up Thanos with S3 backend for long‑term storage (~$5/month)
• Install Loki and Promtail; ship logs (~20 GB/day)
• Deploy Uptime Kuma on the same VM for status page
• Set up SigNoz for distributed tracing of main API
• Configure Alertmanager to route alerts to Discord, email, and PagerDuty

Evidence

Datadog wanted $1,500/month for my setup.

Source

Prometheus scrapes metrics from my services (Node Exporter for hosts, app-specific exporters for apps).

Source

Loki is "Prometheus but for logs" — it doesn't index log content, just labels.

Source

Uptime Kuma ... pings every 5 minutes, alerts via email/Discord/Telegram on failure.

Source

SigNoz is a Datadog APM alternative. It's OpenTelemetry-native, so any OTel-instrumented service works.

Source