SelfHostStackOpen-Source Directory

Why Migrate Away from Datadog APM & New Relic?

Commercial APM platforms like Datadog, New Relic, and Dynatrace employ complex multi-tiered pricing that charges per host, per container, and per gigabyte of ingested trace telemetry. When microservices experience traffic spikes or debug logging is enabled, monthly APM bills can skyrocket unexpectedly. Furthermore, proprietary agent SDKs lock codebases into vendor-specific APIs. Self-hosting SigNoz or Jaeger gives engineering teams OpenTelemetry-native distributed tracing, service dependency maps, exception tracking, metrics, and logs with a lightning-fast ClickHouse columnar database backend — providing complete visibility into production performance with zero per-host or per-gigabyte fees.

Technical Architecture & Migration Analysis

Commercial APM solutions like Datadog require installing proprietary host daemons with closed-source kernel hooks that stream telemetry into proprietary cloud storage with metered per-gigabyte billing. OpenTelemetry-native self-hosted platforms (SigNoz and Jaeger) decouple data collection from storage. Applications instrument standard OTel open-source SDKs that push metrics, traces, and logs directly to an OTel Collector. SigNoz stores telemetry in high-compression ClickHouse columnar tables, enabling instant SQL analysis with zero license fees or ingest limits.

⚠️

When NOT to Migrate (When Staying on Datadog APM & New Relic Makes Sense)

Self-hosting is not universally the right move. Keep paying for SaaS if your team hits any of these constraints:

  • Your company has multi-year enterprise volume commitments with Datadog or New Relic.
  • Your team lacks capacity to manage a 4GB RAM ClickHouse VPS.
  • You require specialized legacy mainframe or proprietary ERP agent plugins not supported by OpenTelemetry.

Real-World Cost Comparison: Datadog APM & New Relic vs Self-Hosted

Comparing vendor cloud billings against standard Hetzner / DigitalOcean infrastructure costs at scale.

Tier / ScaleDatadog APM & New Relic CostSelf-Hosted VPS CostEstimated Annual SavingsTechnical Breakdown
Startup / Dev Team (5 Microservices, 3 Hosts)
5 microservices, 50M trace spans/mo, 20GB log data
$1,860–$3,600/year (Datadog APM Pro $31/host/mo + trace ingest fees)€7.05/month (€84.60/year on Hetzner CPX21)$1,775–$3,515/yearSelf-hosted SigNoz with ClickHouse runs all metrics, traces, and logs on a single VPS.
Mid-Market Engineering (20 Microservices, 15 Hosts)
20 services, 500M spans/mo, distributed tracing + error tracking
$11,160–$24,000/year (Datadog Enterprise APM + New Relic Full Stack)€36.47/month (€437.64/year on Hetzner CPX31)$10,722–$23,562/yearHigh-throughput ClickHouse storage handles billions of traces with sub-second queries.
High-Scale Tech Enterprise (50+ Hosts, High Traffic)
50+ cloud hosts, 5B+ spans/month, real-time distributed latency heatmaps
$60,000–$120,000+/year (Enterprise tier APM contracts)€120.00/month (Dedicated clustered ClickHouse + OTel collectors)$58,560–$118,560+/year100% open-source telemetry with standard OpenTelemetry SDKs.

Top 2 Recommended Open-Source Replacements

Tested, self-contained, and production-ready. Click any tool to inspect verified docker-compose configurations, hardware sizing, and deployment guides.

SigNoz

Apache-2.0⭐ 20.2k+

OpenTelemetry-native APM, distributed tracing, metrics, and logs with ClickHouse columnar storage.

Min RAM4 GB
Min CPU2 vCPU
GitHub Repo ↗

✅ Advantages

  • Direct 1-to-1 open-source alternative to Datadog APM with modern UI
  • Zero proprietary lock-in: applications use standard OpenTelemetry instrumentation
  • ClickHouse columnar storage enables sub-second trace search across billions of spans

⚠️ Trade-offs / Limitations

  • Minimum 4GB RAM required for ClickHouse and OTel collector containers
  • Advanced multi-tenant organization workspaces require enterprise license

Core Features

100% OpenTelemetry native: instrument apps with standard OTel SDKs (Python, Node.js, Go, Java, Rust)
Unified Traces, Metrics & Logs: jump directly from a slow p99 trace span to exact error logs
ClickHouse columnar database engine: 10x-50x faster queries and 10x compression over Elasticsearch
Service Dependency Maps: automated visual dependency graphs with real-time error propagation
Exception & Error Tracking: automated stack trace aggregation with line-level code attribution
Custom Grafana-style dashboards and visual alerting with Slack/PagerDuty/Webhook webhooks

Architecture Notes

Unified observability platform built natively on OpenTelemetry (OTel) collectors and ClickHouse columnar database. Includes a modern React web frontend that unifies distributed traces, p99 latency heatmaps, error rates, infrastructure metrics, and application logs in a single UI.

Known Limitations

ClickHouse requires at least 4GB RAM for stable production write pipelines with high trace ingestion rates.

Official Documentation ↗
📄 docker-compose.yml
Production Ready
# SigNoz standalone docker-compose deployment:
version: '3.8'
services:
  clickhouse:
    image: clickhouse/clickhouse-server:24.1-alpine
    container_name: signoz-clickhouse
    restart: always
    volumes:
      - clickhouse_data:/var/lib/clickhouse
    environment:
      - CLICKHOUSE_DB=signoz_traces
    networks:
      - selfhost_net

  signoz-otel-collector:
    image: signoz/signoz-otel-collector:0.102.0
    container_name: signoz-otel-collector
    restart: always
    ports:
      - "4317:4317" # OTLP gRPC
      - "4318:4318" # OTLP HTTP
    depends_on:
      - clickhouse
    networks:
      - selfhost_net

  signoz-frontend:
    image: signoz/frontend:latest
    container_name: signoz-frontend
    restart: always
    ports:
      - "3301:3301"
    depends_on:
      - clickhouse
    networks:
      - selfhost_net

volumes:
  clickhouse_data:

networks:
  selfhost_net:
    external: true

🚀 5-Minute Deployment Guide

  1. 1Provision a VPS with at least 4GB RAM and 2 vCPUs (e.g. Hetzner CX22/CPX21).
  2. 2Install Docker: `curl -fsSL https://get.docker.com | sh`.
  3. 3Clone official SigNoz deployment repo: `git clone -b main https://github.com/SigNoz/signoz.git && cd signoz/deploy`.
  4. 4Run installation script: `./install.sh`.
  5. 5Open `http://your-server-ip:3301` to register the admin account.
  6. 6Instrument your Node.js/Python/Go applications using standard OpenTelemetry SDKs and send traces to `http://your-server-ip:4318/v1/traces`.

Recommended Cloud VPS for SigNoz

Compare all VPS hosts →
Hetzner Cloud€7.05/mo

CPX21 (3 vCPU, 4GB RAM, 80GB NVMe)

Excellent performance for ClickHouse ingestion of 10M+ spans/day.

Deploy on Hetzner →
DigitalOcean$24.00/mo

General Purpose Droplet (2 vCPU, 4GB RAM, 50GB SSD)

Smooth SSD performance with automated snapshot backups.

Claim $200 DO Credit →

Jaeger Tracing

Apache-2.0⭐ 20.8k+

CNCF graduated open-source end-to-end distributed tracing and transaction monitoring.

Min RAM1 GB
Min CPU1 vCPU
GitHub Repo ↗

✅ Advantages

  • Ultra-lightweight all-in-one container for quick local and staging deployments
  • Rock-solid CNCF open-source governance and enterprise stability
  • Universal compatibility with all OpenTelemetry SDKs

⚠️ Trade-offs / Limitations

  • Specialized strictly in traces (requires separate tools for metrics and logs)
  • Long-term petabyte storage requires external Elasticsearch or OpenSearch cluster

Core Features

CNCF graduated project with immense enterprise adoption (Uber, Red Hat, Apple)
Native OpenTelemetry OTLP ingestion over gRPC (`4317`) and HTTP (`4318`)
Rich trace waterfall visualization with microsecond span timing breakdowns
Dynamic trace sampling and adaptive sampling strategies to control storage volume
Distributed context propagation for tracing requests across polyglot microservice architectures
Integrates seamlessly with Grafana Tempo and Prometheus monitoring stacks

Architecture Notes

Graduated Cloud Native Computing Foundation (CNCF) distributed tracing platform. Includes an OpenTelemetry-compatible trace collector, query engine, in-memory/Elasticsearch/OpenSearch/Cassandra storage backends, and a battle-tested trace visualization web UI.

Known Limitations

Jaeger focuses strictly on distributed tracing and span waterfalls; does not include built-in log aggregation or server metric charts (pairs well with Prometheus + Grafana).

Official Documentation ↗
📄 docker-compose.yml
Production Ready
version: '3.8'
services:
  jaeger:
    image: jaegertracing/all-in-one:latest
    container_name: jaeger
    restart: always
    ports:
      - "16686:16686" # Jaeger UI
      - "4317:4317"   # OTLP gRPC
      - "4318:4318"   # OTLP HTTP
    environment:
      - COLLECTOR_OTLP_ENABLED=true
    networks:
      - selfhost_net

networks:
  selfhost_net:
    external: true

🚀 5-Minute Deployment Guide

  1. 1Provision a lightweight Linux VPS.
  2. 2Install Docker: `curl -fsSL https://get.docker.com | sh`.
  3. 3Run Jaeger all-in-one container: `docker run -d --name jaeger -p 16686:16686 -p 4317:4317 -p 4318:4318 -e COLLECTOR_OTLP_ENABLED=true jaegertracing/all-in-one:latest`.
  4. 4Access Jaeger Web UI at `http://your-server-ip:16686`.
  5. 5Configure your application's OpenTelemetry exporter to send traces to `http://your-server-ip:4318`.
  6. 6Visualize distributed trace waterfalls and diagnose backend query bottlenecks.

Recommended Cloud VPS for Jaeger Tracing

Compare all VPS hosts →
Hetzner Cloud€3.79/mo

CX22 (2 vCPU, 4GB RAM, 40GB NVMe)

Perfect server for Jaeger all-in-one and Prometheus monitoring.

Deploy on Hetzner →

Quick Specification Matrix

ToolLicenseMin RAMMin CPUGitHub RepoPrimary Advantage
Datadog APM & New Relic (Proprietary)Proprietary ClosedManaged CloudManaged CloudN/ATurnkey onboarding with vendor lock-in & paywalls
SigNozApache-2.04 GB2 vCPUSigNoz/signozDirect 1-to-1 open-source alternative to Datadog APM with modern UI
Jaeger TracingApache-2.01 GB1 vCPUjaegertracing/jaegerUltra-lightweight all-in-one container for quick local and staging deployments

Performance Benchmarks & Hard Operational Limits

Real-world operational trade-offs, resource consumption limits, and measured throughput.

Benchmark MetricDatadog APM & New Relic BaselineSelf-Hosted Alternative MetricOperational Bottleneck / LimitSource
Telemetry Storage Compression RatioN/A (Cloud metered by uncompressed GB ingest)8x to 15x columnar compression with ClickHouse (SigNoz)NVMe disk write speed.Production Test
Trace Query Latency (p99 latency over 100M spans)1.5s–4.0s (Datadog cloud web interface)0.2s–0.8s (Native vectorized ClickHouse SQL query)Memory and CPU cores.Production Test
Per-Host & Per-GB Ingest Fees$31-$65/host/mo + $0.10/GB trace ingest$0.00 (Flat server hosting cost only)None.Production Test

Frequently Asked Questions

Practical deployment, migration, and maintenance answers.

Can I use standard OpenTelemetry SDKs without changing code when migrating to SigNoz?

Yes. SigNoz is 100% native OpenTelemetry. If your application already uses standard OTel instrumentation, you simply set `OTEL_EXPORTER_OTLP_ENDPOINT=http://your-signoz-collector:4318` without altering your application code.

How does ClickHouse make SigNoz so fast and resource-efficient?

ClickHouse is a columnar database designed specifically for real-time analytical queries. It compresses trace and log data by up to 15x and executes vectorized multi-core queries across billions of telemetry records in milliseconds.

What is the primary difference between SigNoz and Jaeger?

Jaeger focuses strictly on distributed tracing and span visualization. SigNoz provides an all-in-one observability suite combining distributed tracing, metrics, application logs, and customizable Grafana-style dashboards.

How much RAM does SigNoz require in production?

SigNoz runs stably on 4GB RAM (e.g. Hetzner CPX21 for €7.05/mo) for small-to-medium workloads. For high-volume production systems ingesting tens of millions of spans daily, 8GB to 16GB RAM is recommended.

Does SigNoz support alerting for latency spikes and error rate thresholds?

Yes. SigNoz includes a visual alert builder where you can define threshold-based alerts (e.g. 'Trigger if p99 latency > 500ms for 5 minutes') and send notifications to Slack, PagerDuty, or custom webhooks.

Starter Stack Pack — $29

Skip the setup: get the production-ready stack

Don't stitch together configs from five different READMEs. Get all 5 production-hardened Docker Compose stacks — Postgres, Redis, SSL auto-renewal, and backup scripts — ready to deploy in minutes.

n8nVisual workflow automation
📊UmamiPrivacy-first web analytics
🛡️Uptime KumaUptime monitoring & alerts
🔐VaultwardenBitwarden-compatible vault
☁️NextcloudDropbox/Drive replacement
Get the Stack Pack — $29 →

One-time purchase · Instant download · Production-ready

esc
navigate open