Choosing among open source APM tools in 2026 starts with one question: are you picking instrumentation, a trace backend, a composable stack, or a unified platform? This guide walks through ten names teams actually encounter when searching opensource apm â from OpenTelemetry and Jaeger through Grafana LGTM, Elastic APM, SigNoz, ClickHouse-backed analytics, and OTel-native unified backends like DataBuff â with selection criteria you can apply without copying vendor ranking lists.
Application performance monitoring under an open source license still wins when teams need data ownership, portable instrumentation, and the freedom to self-host without procurement friction. The through-line across every entry below is OpenTelemetry: most modern backends ingest OTLP, and the real choice is whether a tool is OTel-native from the ground up or bolts OTLP onto an older data model.
Top open source APM tools
Here are ten projects and patterns that dominate open source apm tools shortlists in 2026. The list mixes standards, trace specialists, full-stack platforms, composable stacks, storage engines, and unified backends â because production teams rarely pick just one layer.
1. OpenTelemetry
OpenTelemetry is not an APM product â it is the CNCF-backed standard for generating and exporting telemetry. SDKs, auto-instrumentation libraries, and the OpenTelemetry Collector form the instrumentation layer that every backend below consumes.
OpenTelemetry · SDKs, Collector, and OTLP export to observability backends
OpenTelemetry provides vendor-neutral APIs for traces, metrics, and logs; semantic conventions for HTTP, database, and messaging spans; and a Collector pipeline for sampling, routing, and fan-out to multiple destinations.
Key strength: Industry-standard instrumentation that decouples how you observe code from where you store telemetry â the foundation every modern OSS APM stack builds on.
2. Jaeger
Jaeger is a distributed tracing system originally built at Uber and now a CNCF graduated project. Jaeger v2 re-architected around the OpenTelemetry Collector framework, aligning ingest with modern OTLP deployments while preserving deep trace search, adaptive sampling, and service dependency graphs.
Jaeger UP· Trace search and service filtering
Jaeger focuses on the tracing pillar of APM. Teams typically pair Jaeger with Prometheus for metrics and Grafana for dashboards when they need full-stack observability outside a single product UI.
Key strength: Mature, CNCF-aligned distributed tracing with native OTLP ingestion and flexible pluggable storage.
3. DataBuff
DataBuff is an open source, AI-native APM platform built for the OpenTelemetry era. Listed on the OpenTelemetry Vendors page under Pure OSS with Native OTLP Yes, it ingests standard OTLP directly from applications and the Collector.
DataBuff · Service health overview (Rate, Errors, Duration)
Architecturally, DataBuff collapses observability sprawl into three runtime components: Ingest (OTLP gRPC 4317 and HTTP 4318), Apache Doris columnar storage, and a Web platform on port 27403 for service health, topology, alerting, and AI-assisted investigation.
curl -fsSL https://databuff.ai/databuff/ai-apm-install.sh | bash
# OTLP HTTP: http://YOUR_HOST:4318/v1/traces
Key strength: Compact, OpenTelemetry-native unified APM with columnar analytics storage, AI-native multi-agent triage, and MCP/skill extensibility for agent-era on-call workflows.
4. Apache SkyWalking
Apache SkyWalking is an open source APM system designed for microservices, cloud-native, and container-based architectures. It provides distributed tracing, service mesh telemetry analysis, metric aggregation, and topology visualization â with agents for Java, .NET, Node.js, Python, and other runtimes.
Apache SkyWalking · Default observability dashboard
SkyWalking predates the OpenTelemetry wave but now exposes OTLP receivers, letting teams migrate instrumentation gradually.
Key strength: Integrated APM for microservices and cloud-native environments with rich agent-based auto-instrumentation and topology maps.
5. Zipkin
Zipkin is one of the earliest open source distributed tracing systems. It collects timing data from services, stores spans, and provides a UI to look up traces and understand request paths and latencies.
Zipkin Lens · Distributed trace lookup
Zipkin's architecture is deliberately simple and focused. Like Jaeger, it handles tracing and relies on companion tools for metrics and logs.
Key strength: A mature, simple system dedicated to distributed tracing with a long track record in production microservice shops.
6. Grafana
Grafana is the visualization layer at the center of the LGTM stack â Loki for logs, Grafana for dashboards, Tempo for traces, and Mimir or Prometheus for metrics. Many teams treat this composable pattern as their open source observability platform.
Grafana + Tempo · TraceQL search in Explore
Trace-to-log and trace-to-metrics correlation work well when Loki, Tempo, and Prometheus are wired together. The trade-off is operational complexity: you operate and upgrade each component separately.
Key strength: Highly flexible dashboards and visualization that integrate with almost any telemetry backend â the de facto standard for composable OSS observability.
7. Elastic APM
Elastic APM is the application performance monitoring component of the Elastic Stack. It collects performance metrics, errors, and distributed traces, storing them in Elasticsearch and visualizing them in Kibana alongside log data.
Elastic APM · Service performance in Kibana
Teams already running Elasticsearch for logging often add Elastic APM for a unified search and analytics experience.
Key strength: Seamless APM integration for teams already standardized on the Elastic Stack for logging and search.
8. Pinpoint
Pinpoint is an open source APM tool inspired by Google Dapper, designed for large-scale distributed systems. Written primarily for Java and PHP, it uses bytecode instrumentation to trace requests without modifying application source code.
Pinpoint · Server map and service dependency topology
Pinpoint visualizes system topology and provides code-level visibility â expandable call trees showing which methods, SQL statements, and external APIs consumed latency inside a transaction.
Key strength: Deep code-level transaction tracing and topology maps for Java and PHP applications at scale.
9. SigNoz
SigNoz is a widely referenced open source observability platform that provides a unified backend for metrics, traces, and logs. It is built natively on OpenTelemetry and stores telemetry in a columnar analytics engine.
SigNoz · Application overview (RPS, latency percentiles, error rate)
SigNoz offers application overview dashboards, log management, infrastructure monitoring, service maps, and alerting in one UI.
Key strength: OpenTelemetry-native unified observability with columnar storage optimized for correlated metrics, traces, and logs.
10. ClickHouse
ClickHouse is not an APM UI â it is an open source columnar OLAP database that powers the storage layer behind several OpenTelemetry-native observability platforms. High-cardinality trace and metrics data benefits from columnar compression and fast analytical queries.
ClickHouse · SQL analytics on high-cardinality telemetry data
When teams mention "ClickHouse APM," they usually mean an observability backend that ingests OTLP into ClickHouse tables and exposes SQL or a product UI on top.
Key strength: High-performance columnar storage for telemetry analytics â the engine behind multiple OTel-native APM backends.
Top open source APM tools at a glance
Read the list by role, not rank:
- OpenTelemetry â Instrumentation standard; defines OTLP; portable telemetry export
- Jaeger â Trace backend; traces only; OTLP-native
- DataBuff â Unified platform; metrics + traces; OTLP-native; AI-native APM
- SkyWalking â Full-stack APM; metrics + traces; OTLP receiver
- Zipkin â Trace backend; traces only; legacy OTLP support
- Grafana â LGTM composable stack; metrics via Mimir/Prom, traces via Tempo, logs via Loki
- Elastic APM â Full-stack APM for ELK-centric estates
- Pinpoint â Java/PHP APM with bytecode-level tracing
- SigNoz â OTel-native unified platform; metrics + traces + logs
- ClickHouse â Columnar storage engine for telemetry analytics
OpenTelemetry and ClickHouse are infrastructure layers. Jaeger and Zipkin are trace specialists. Grafana, SkyWalking, Elastic APM, SigNoz, and DataBuff compete for "where on-call lives."
Choosing the right open source APM tool
- Trace-only pain â Start with Jaeger or Zipkin; add Prometheus and Grafana if RED metrics matter.
- Java-heavy agent preference â Evaluate SkyWalking or Pinpoint for bytecode depth; plan OTLP receivers for new services.
- Existing Grafana practice â Extend with Tempo and Loki rather than replacing the stack on day one.
- ELK-standardized logging â Elastic APM keeps search and APM in one vendor ecosystem.
- Greenfield OpenTelemetry â Shortlist OTel-native unified backends (SigNoz, DataBuff) and validate with the same OTLP POC script.
- AI-assisted on-call â Weight backends that ground LLM workflows in live spans and expose MCP or skill hooks for IDE agents.
Validation script: same steps for every finalist
- Deploy with the project's published install path.
- Point an OpenTelemetry SDK or Collector exporter at OTLP gRPC
4317or HTTP4318. - Generate traffic for five to ten minutes.
- Confirm services appear, dependency topology renders, and a slow trace is searchable end-to-end.
- Record retention defaults, sampling behavior, and resource usage on your target VM class.
For DataBuff, open the web UI on port 27403 after install. Full eBPF zero-instrumentation coverage remains on the public roadmap â plan SDK or Collector instrumentation for brownfield services today.
FAQ
What is the best open source APM tool in 2026?
There is no single best tool. Jaeger excels at dedicated tracing; Grafana LGTM wins for composable stacks; SkyWalking and Pinpoint serve agent-driven Java fleets; SigNoz and DataBuff target OTel-native unified platforms. Match the category to your signal mix and ops capacity.
Is OpenTelemetry an APM tool?
No. OpenTelemetry is the instrumentation and export standard. APM tools are backendsPand UIs that receive OTLP and provide service health, trace search, and alerting workflows.
Why include ClickHouse in an APM list?
Because several OTel-native APM platforms store telemetry in ClickHouse or similar columnar engines. Understanding that layer helps you plan retention, cardinality, and query cost.
Can open source APM replace Datadog or New Relic?
Modern OSS platforms offer comparable tracing, metrics, logs, and alerting for many workloads. The trade-off is self-hosting operations versus managed SaaS convenience.
How does DataBuff differ from SigNoz?
Both ingest OTLP into columnar storage for unified observability. DataBuff emphasizes a three-component footprint (Ingest, Apache Doris, Web), AI-native multi-agent triage grounded in live telemetry, and Skill/MCP extensibility for IDE agent workflows.
Closing perspective
The open source apm tools landscape in 2026 is healthier because OpenTelemetry decoupled instrumentation from backend choice. Use this top-ten map to identify which layer you are actually selecting â standard, trace specialist, composable stack, storage engine, or unified platform â then run the same OTLP validation script on every finalist.
References: OpenTelemetry · OTel Vendors · Jaeger · SkyWalking · DataBuff
Top comments (0)