Datadog vs Uptrace: APM & Observability Comparison

Compare Datadog and Uptrace on pricing, features, and fit — including why small teams switch from Datadog's per-host billing to Uptrace's open-source, OpenTelemetry-based tracing.

Updated 2026-09 · 2026

Datadog

All-in-one monitoring, APM, logs, and security platform

$15per host/month (Infrastructure Pro); APM adds $31/host/month

Strengths

+Huge integration catalog (750+) covering nearly every cloud service and tool
+Single platform for infra metrics, APM, logs, RUM, and security in one UI
+Mature alerting, dashboards, and anomaly detection out of the box

Weaknesses

-Billing is complex and stacks up fast — hosts, APM, logs, custom metrics, and indexed spans are all billed separately
-Costs scale unpredictably; surprise bills are a common complaint among small teams
-Vendor lock-in via proprietary agents and query language

Best for

Funded companies that need a fully managed, all-in-one observability suite and can absorb per-host/per-GB costs at scale.

Uptrace

Open source APM and distributed tracing built on OpenTelemetry

Freeself-hosted (open source); usage-based cloud plans available

Strengths

+Fully open source (BSL/Apache-licensed components) — self-host at zero cost
+Native OpenTelemetry support, no proprietary agent to install or lock into
+Traces, metrics, and logs correlated in one query language (UQL)

Weaknesses

-Smaller ecosystem and fewer pre-built integrations than Datadog
-You own the operational burden of running ClickHouse and Uptrace yourself if self-hosting
-No built-in security monitoring, RUM, or synthetic testing like Datadog offers

Best for

Engineering teams already using OpenTelemetry who want tracing and APM without per-host or per-GB SaaS pricing.

Feature Comparison

Feature	Datadog	Uptrace
Distributed tracing	Yes, via Datadog APM agent or OTel	Yes, native OpenTelemetry-first design
Log management	Yes, billed separately by GB ingested/indexed	Yes, included, stored in ClickHouse
Infrastructure metrics	Yes, extensive, agent-based	Yes, via OTel collector
Self-hosting option	No, SaaS only	Yes, fully self-hostable for free
Pricing model	Per host + per GB + per feature add-on	Free self-hosted, or usage-based cloud
OpenTelemetry native	Supported via OTel ingest, but proprietary agent preferred	Built OTel-first from the ground up
Alerting & dashboards	Mature, highly customizable	Basic built-in dashboards and alerts
Security monitoring (SIEM/CSPM)	Yes, dedicated products	No
Real User Monitoring (RUM)	Yes	No
Integration count	750+	Anything emitting OTel data
Free tier	5 hosts, 1-day retention	Unlimited when self-hosted
Vendor lock-in risk	High (proprietary agent/format)	Low (standard OTel data)

The Verdict

Datadog wins if you want a single managed platform that does everything and money isn't the constraint — but its per-host, per-feature billing punishes small teams as they grow. Uptrace is the better fit for teams already on OpenTelemetry who want tracing and APM without a scaling SaaS bill, at the cost of running more infrastructure themselves. If your Datadog bill is dominated by APM/host charges rather than security or RUM features, Uptrace is a serious downgrade path in cost, not capability.

How to switch from Datadog to Uptrace

1Export your existing Datadog monitor and dashboard configs as JSON using the Datadog API (/api/v1/monitor and /api/v1/dashboard endpoints) so you have a reference to rebuild them.
2Stand up Uptrace — either self-hosted via Docker Compose (Uptrace + ClickHouse) or sign up for Uptrace Cloud — and grab your project's DSN/API key.
3Configure an OpenTelemetry Collector to send traces, metrics, and logs to Uptrace's OTLP endpoint; run it in parallel with the Datadog agent during the transition.
4Recreate your critical alerts and dashboards in Uptrace using its query language (UQL), starting with the top 5-10 monitors your team actually acts on.
5Run both Datadog and Uptrace side by side for 1-2 weeks to validate trace/log parity, then redirect application instrumentation fully to the OTel collector pointing at Uptrace.
6Downgrade or cancel the Datadog plan once your team confirms Uptrace covers on-call needs, and archive old Datadog dashboards/monitors for reference.

Datadog vs Uptrace: common questions

How do I export my data from Datadog before switching?+

Datadog doesn't offer a bulk 'export everything' button — you pull data out via the API (metrics, logs, and traces endpoints) or set up a dual-write period where both Datadog and Uptrace receive OpenTelemetry data simultaneously. Most teams export dashboard/monitor definitions as JSON via the Datadog API and rebuild them manually in Uptrace since the underlying data models differ.

What do I lose by moving from Datadog to Uptrace?+

You lose Datadog's RUM, Security Monitoring (SIEM/CSPM), Synthetic Monitoring, and the huge pre-built integration library. You keep core APM, distributed tracing, logs, and metrics, but expect to rebuild dashboards and alerts manually since Uptrace uses its own query language (UQL) instead of Datadog's.

Is Uptrace's free self-hosted version enough for a small team?+

Yes, for most small teams — self-hosted Uptrace on a single ClickHouse instance easily handles moderate trace/log volume for free, aside from your own server costs. If you don't want to manage infrastructure, Uptrace also offers a paid cloud-hosted option with usage-based billing instead of per-host pricing.

Does Uptrace integrate with the same tools as Datadog?+

Uptrace integrates with anything that emits OpenTelemetry data — which covers most modern languages, frameworks, and infrastructure tools — but it doesn't have Datadog's 750+ pre-built dashboards and one-click integrations. You'll typically configure an OTel collector once and point it at Uptrace instead of using vendor-specific agents.

How does the cost actually compare over a year?+

A team paying Datadog $31/host/month for APM plus log/metric overages can easily hit $1,000+/month at 20-30 hosts, while self-hosted Uptrace costs only your ClickHouse server (often under $100/month on cloud infra). Even Uptrace's paid cloud tier is usage-based rather than per-host, so costs grow with actual data volume, not headcount or server count.