Datadog alternatives 2026: an honest comparison of the most important options
Nine Datadog alternatives compared fairly: pricing models, OpenTelemetry approach, hosting options and who each one fits best, plus a migration checklist you can copy.
Summary
- Teams look for Datadog alternatives mainly because of cost predictability, instrumentation lock-in and data residency, not because Datadog lacks features.
- The nine options below differ most in their pricing unit, how natively they handle OpenTelemetry and where your data can live, so compare on those axes first.
- Whatever you pick, migrate via OpenTelemetry with a parallel-run phase, so your instrumentation stays portable and the next switch is cheaper.
Searching for Datadog alternatives usually starts with an invoice or a contract renewal, not with a missing feature. Datadog is a broad, mature platform, and many teams run it happily. Others find that per-host and per-product pricing becomes hard to forecast, that proprietary agents and libraries make a switch expensive, or that they need tighter control over where telemetry is stored.
Disclosure: this article is published by Obseria, which is one of the options compared below. We have tried to describe every tool fairly, based on the vendors' own public pricing pages and documentation as of October 2026. Prices and features change, so verify the details that matter to you before you decide.
Below: why teams evaluate a Datadog alternative, the criteria that matter, nine short profiles, a comparison table and a migration checklist.
Why teams look for Datadog alternatives
In our conversations with platform teams, three reasons come up again and again. None of them means Datadog is a bad product; they describe a mismatch between a pricing and deployment model and a team's situation.
- Cost predictability: Datadog bills infrastructure and APM per host, plus per-GB and per-event units for logs and other products, each with its own SKU. With autoscaling, short-lived containers and growing log volumes, forecasting the monthly bill takes effort. Datadog offers tools such as Observability Pipelines and Flex Logs for this.
- Lock-in: the Datadog Agent and tracing libraries work well, but code instrumented with vendor-specific libraries has to be re-instrumented when you leave. Datadog supports OTLP ingest via its Agent and its own OpenTelemetry Collector distribution, which helps, yet many existing estates still run on proprietary instrumentation. See our glossary entry on vendor lock-in for the general pattern.
- Data residency and sovereignty: Datadog runs an EU1 site in Germany, among other regional sites. For some organizations that is enough; others want a European provider, a bring-your-own-cloud (BYOC) deployment or a fully self-hosted setup for regulatory or procurement reasons.
Whichever is your trigger should carry the most weight in your evaluation.
Evaluation criteria that actually matter
Almost every serious alternative covers metrics, logs and traces. The differences that decide whether a switch pays off are elsewhere:
- Pricing unit: per host, per GB, per event or data point, per user or committed capacity. Model your real volumes, today and at three times today, in each pricing model.
- OpenTelemetry approach: is OTLP the native ingest path, an additional endpoint or a vendor distribution? Can you export your data in an open format?
- Hosting and residency: SaaS regions, BYOC, managed or self-hosted, and who operates what.
- Feature packaging: are APM, RUM, synthetics, SLOs and incident management included or separate products?
- Operational effort: a self-hosted stack saves licence fees but costs engineering time for upgrades, scaling and on-call.
- Migration path: dashboards, monitors and alert rules must be rebuilt somewhere; ask what tooling and support the vendor offers.
For a full, vendor-neutral scoring sheet, use our guide on how to choose an observability platform.
The most important Datadog alternatives at a glance
The profiles are listed with Obseria first because it is our product; the remaining order carries no ranking.
1. Obseria
An OpenTelemetry-native platform from Just Growing GmbH (Germany). Telemetry passes through Obseria Pipeline (filtering, PII masking, sampling, routing before storage) into Obseria Lake, an index-free store for logs, metrics, traces and events. Pricing is per million events or data points ingested and stored, with unlimited users and all features included. Hosting: SaaS in the EU (Frankfurt, Dublin) or UK (London), or BYOC.
Good fit if you want EU data residency, OTLP in and out, and one usage-based price without per-host or per-user fees. Consider that Obseria is a young company founded in 2025 and has a smaller ecosystem of community integrations than long-established vendors. Details: Obseria vs Datadog.
2. Dynatrace
A full-stack platform built around OneAgent auto-instrumentation, with a native OTLP endpoint and the Grail data store. Commercially it uses the Dynatrace Platform Subscription: a minimum annual commitment drawn down by a rate card (for example host-hours, GiB ingested and queried). It runs as SaaS on AWS, Azure and Google Cloud, or as Dynatrace Managed operated by the customer.
Good fit if you are a large enterprise that values deep automatic instrumentation and AI-driven analysis and can plan an annual commitment. Consider the commitment model and whether OneAgent or OpenTelemetry will be your primary instrumentation. Details: Obseria vs Dynatrace.
3. New Relic
A broad SaaS platform priced per GB ingested plus user licences (or compute capacity units), with a perpetual free tier of 100 GB ingest per month and one full platform user. OTLP is the recommended ingest path for OpenTelemetry data, with US and EU endpoints; EU data storage carries a surcharge per GB.
Good fit if a small number of engineers need full access and you like simple per-GB data pricing. Consider how many full platform users you need, because user licences can dominate the bill in larger teams. Details: Obseria vs New Relic.
4. Grafana Cloud / LGTM stack
The managed version of the open-source Loki, Grafana, Tempo and Mimir stack, with native OTLP ingest and Grafana Alloy as an OpenTelemetry Collector distribution. Pricing combines a platform fee on Pro with usage per active series, GB and product unit, plus some per-user products. Available as Grafana Cloud (including a Frankfurt region), BYOC on Enterprise, or self-hosted open source under AGPLv3.
Good fit if your team already knows Prometheus and Grafana, wants the option to self-host and is comfortable with one query language per signal. Consider that separate backends per signal and separately billed products mean more moving parts to understand. Details: Obseria vs Grafana.
5. Elastic Observability
Built on Elasticsearch, with the Elastic Distribution of OpenTelemetry (EDOT) as well as Elastic Agent and APM agents. Serverless is priced per GB ingested and retained; Elastic Cloud Hosted is resource-based; self-managed deployments, including on-premises, are licensed by nodes or RAM. Frankfurt regions are available on AWS, Azure and Google Cloud.
Good fit if log search is your main use case, you already run Elastic for search or security, or you need an on-premises option. Consider the operational work of index lifecycle management in hosted and self-managed setups. Details: Obseria vs Elastic.
6. Honeycomb
A tracing-first platform known for high-cardinality event analysis and BubbleUp. Plans (Free, Pro, Enterprise) come with event allowances; the free plan includes up to 20 million events per month. OTLP is the recommended ingest path. Data lives in the US or in the EU (AWS Ireland).
Good fit if debugging distributed systems through traces is your core need and your engineers like exploratory querying. Consider that some capabilities such as the service map or frontend observability are tied to the Enterprise plan. Details: Obseria vs Honeycomb.
7. Dash0
An OpenTelemetry-native SaaS built on the OpenTelemetry Collector, with per-signal pricing for ingest and storage, PromQL and SQL querying and the Agent0 AI agent. It runs in EU or US regions on AWS and Google Cloud and can be bought via several cloud marketplaces.
Good fit if you want a lean, OTel-first SaaS with transparent per-signal pricing and resources manageable as code. Consider that the product focuses on core observability, so check whether you need additional modules like incident management or status pages elsewhere. Details: Obseria vs Dash0.
8. SigNoz (open source)
An open-source, OpenTelemetry-native platform for logs, metrics and traces in a single columnar database. You can self-host the Community Edition (Docker, Linux or Kubernetes via Helm) or use SigNoz Cloud, which lists data centers in the US, EU and India. The cloud Teams plan starts at $49 per month including $49 of usage, with logs and traces billed per GB and metrics per million samples; Enterprise adds dedicated cloud, BYOC or supported self-hosting.
Good fit if you want an open-source, OTel-native alternative with a single UI and the choice between self-hosting and SaaS. Consider that some features, such as SSO/SAML and anomaly detection, are not part of the Community Edition.
9. Self-hosted Prometheus + OpenTelemetry stack
The do-it-yourself route: OpenTelemetry SDKs and the Collector for instrumentation, Prometheus for metrics and open-source backends for logs and traces, visualized with Grafana. Both Prometheus and OpenTelemetry are CNCF graduated projects; OpenTelemetry reached graduated status in May 2026.
Good fit if you have strict sovereignty requirements, a strong platform team and want zero licence fees. Consider that you pay in engineering time: high availability, long-term storage, upgrades, scaling and the on-call duty for the monitoring system itself.
Comparison table: Datadog alternatives side by side
| Tool | Pricing model | OpenTelemetry approach | Hosting options | Best for |
|---|---|---|---|---|
| Datadog (baseline) | Per host plus per-GB and per-event units by product | OTLP via Agent and Datadog OTel Collector distribution | SaaS on regional sites incl. EU1 | Teams wanting one broad commercial suite |
| Obseria | Per million events or data points; unlimited users | OTel-native, OTLP in and export | SaaS EU/UK or BYOC | EU residency with usage-based pricing |
| Dynatrace | Annual commitment drawn down by rate card | OneAgent plus native OTLP endpoint | SaaS or Dynatrace Managed | Large enterprises, deep auto-instrumentation |
| New Relic | Per GB ingested plus user licences | Native OTLP, US and EU endpoints | SaaS in US or EU | Small teams with few full users |
| Grafana Cloud | Platform fee plus usage per series, GB, unit; some per user | Native OTLP; Alloy collector | SaaS, BYOC (Enterprise), self-hosted OSS | Prometheus and Grafana users |
| Elastic | Per GB (Serverless), resources (Hosted) or nodes (self-managed) | EDOT Collector and SDKs | Serverless, Hosted, self-managed incl. on-prem | Log-heavy and on-premises needs |
| Honeycomb | Plans with event allowances | OTLP recommended ingest path | SaaS US or EU | Trace-driven debugging |
| Dash0 | Per signal, ingest and storage | OTel-native, built on Collector | SaaS EU or US | Lean OTel-first SaaS |
| SigNoz | Cloud: base plan plus per GB / per million samples; OSS free | OTel-native | Cloud, BYOC, self-hosted | Open-source preference |
| Prometheus + OTel | No licence; infrastructure and staff costs | OpenTelemetry by definition | Self-hosted only | Strong platform teams, full control |
To compare the pricing models with your own volumes, our cost calculator lets you plug in hosts, log GB and spans per month. Treat the result as an estimate.
How to shortlist two or three candidates
Start from your trigger. If cost predictability is the problem, shortlist tools whose pricing unit maps to something you can forecast, and model a scaling scenario. If lock-in is the problem, require OTLP as the native path and test data export. If residency is the problem, ask for the concrete region, the legal entity of the processor and whether BYOC or self-hosting is possible.
Then run a two to four week proof of concept on your own telemetry with two or three vendors. Mirror the same OpenTelemetry Collector output to all of them, so you compare platforms, not instrumentation. Measure incident query speed, dashboard rebuild effort and projected monthly cost.
Migration checklist: moving away from Datadog
Copy this checklist into your ticket system. It assumes OpenTelemetry as the target instrumentation, which keeps you portable whichever alternative you choose; our guide to migrating from proprietary agents to OpenTelemetry covers the technical steps in depth.
[ ] 1. Inventory: services, hosts, containers, log GB/day, spans/day, custom metrics
[ ] 2. Export dashboards, monitors and SLO definitions (API/JSON) and rank by importance
[ ] 3. Note contract end date and notice period; plan overlap budget
[ ] 4. Deploy OpenTelemetry Collector; send OTLP to Datadog AND the new platform
[ ] 5. Replace vendor tracing libraries with OTel SDKs, service by service
[ ] 6. Add filtering, PII masking and sampling in the pipeline before storage
[ ] 7. Rebuild top dashboards and alerts; compare alert behaviour for 2+ weeks
[ ] 8. Map RUM, synthetics, on-call and status pages to the new setup
[ ] 9. Train on-call engineers; update runbooks and escalation links
[ ] 10. Cut over alerting, stop dual shipping, decommission agents
[ ] 11. Archive or export historical data you must retain
[ ] 12. Review cost and incident metrics 30 and 90 days after cut-overDual shipping is the safest pattern: for a few weeks both platforms receive identical data, so you can compare alerts and dashboards before anyone loses visibility.
Common mistakes when switching
- Comparing list prices without your volumes: per-host, per-GB and per-event prices cannot be compared directly. Always model your own numbers.
- Rebuilding every dashboard: many dashboards are unused. Migrate the ones people opened in the last 90 days and delete the rest.
- Swapping one proprietary agent for another: if the new platform also requires vendor-specific instrumentation, the next migration will be just as costly.
How Obseria helps
Obseria is built for teams that want to keep OpenTelemetry as their standard and need predictable, usage-based pricing with EU data residency. Because ingest is plain OTLP, you can dual ship from your existing Collector during a proof of concept without touching application code.
- Pricing per million events or data points, unlimited users, all features included.
- Obseria Pipeline filters, masks PII and samples before storage, so you pay only for data you keep.
- SaaS in Frankfurt, Dublin or London, or BYOC in your own cloud account.
- A structured migration service and a switch bonus to cover overlapping contract periods.
If Obseria is on your shortlist, start a free 14-day trial and point your OpenTelemetry Collector at it alongside your current platform.