Back to catalog
Uptrace
AGPL-3.0Uptrace is an open-source APM backend ingesting OpenTelemetry traces, metrics, and logs into ClickHouse to replace costly per-host Datadog licensing.
Open-source alternative to — compare it against:
Overview
Uptrace is an open-source application performance monitoring backend designed to ingest, correlate, and visualize OpenTelemetry traces, metrics, and logs. It couples standard OTLP ingestion directly with a ClickHouse columnar datastore to query distributed telemetry without proprietary collectors or per-host license meters.
Why Uptrace instead of Datadog APM
Datadog meters APM hosts, custom metrics, and ingested traces separately, driving unpredictable monthly costs that force teams to throttle trace sampling. Adapting to its ecosystem often requires proprietary agent wrappers and specific tagging schemas. In contrast, Uptrace relies entirely on standard OpenTelemetry Protocol (OTLP) ingestion. Developers instrument services with standard language SDKs and export data through an OpenTelemetry Collector, eliminating proprietary agent daemons and ensuring telemetry pipelines remain portable across backends.
For daily triage, Uptrace unifies distributed traces, logs, and runtime metrics, allowing developers to jump from high-latency spikes directly into associated span metadata. It also acts as a Tempo and Prometheus datasource for teams maintaining existing Grafana boards. The compromise lies in operational scope. Datadog provides managed continuous profiling, synthetic monitoring, and cloud security features. Uptrace strictly handles core APM; it eliminates per-host licensing and grants full data control, but requires platform teams to manage ClickHouse storage and Postgres operations.
What it does well
- Standard OTLP ingestion: accepts traces, metrics, and logs over gRPC and HTTP without requiring proprietary host agents, keeping application instrumentation vendor-neutral.
- Unified telemetry correlation: links distributed traces directly to structured logs and runtime metrics, accelerating root-cause analysis during service degradation.
- Automated service dashboarding: generates dependency maps, throughput rates, and error charts as soon as services emit spans, removing manual dashboard setup.
- Dual Grafana compatibility: acts as a Tempo and Prometheus datasource, enabling DevOps engineers to retain established visualization workflows.
- Compact ClickHouse storage: compresses telemetry down to roughly 40 bytes per span on disk, making multi-week trace retention cost-effective on modest hardware.