DevStackDecider

Observability

Best APM tools: choose by stack, hosting model and OpenTelemetry approach

APM tools compared on instrumentation, OpenTelemetry, hosting, data regions and billing model, with a shortlist by situation rather than one winner.

DevStackDecider Editorial · 8 Nov 2026 · Vendor documentation

Observability Capability Matrix Products on this page 9 products · verified 8 Nov 2026 – 9 Nov 2026
ProductAPMDistributed tracingMetricsInfrastructure monitoringContinuous profilingReal user monitoringOpenTelemetryRuns asData regionsFree optionPricing modelCompliance on vendor page
DatadogDatadogsource · 8 Nov 2026 YesYesYesYesYesYes Native OTLP ingest SaaS US, EU, UK, APACNine sites incl. EU1 (Germany), UK1, AP1 (Japan), AP2 (Australia) and two US government sites Free tier per host, per GB ingested, per event, per session, per test run SOC 2 Type II, ISO 27001, HIPAA, FedRAMP, PCI DSS
New RelicNew Relicsource · 8 Nov 2026 YesYesYesYesNot statedYes Native OTLP ingest SaaS US, EU, APACUS (default), EU and Japan data-centre regions Free tier per GB ingested, per user, usage credits Not confirmed in sources checked
DynatraceDynatracesource · 9 Nov 2026 YesYesYesYesNot statedYes Native OTLP ingest SaaS, Self-hosted Not statedSaaS regions not listed on the pages checked Trial per host, per GB ingested, per GB retained, per session, usage credits SOC 2 Type II, ISO 27001, HIPAA, FedRAMP
Grafana LabsGrafana Cloudsource · 8 Nov 2026 YesYesYesYesYesYes Native OTLP ingest SaaS US, EU, UK, APAC, OTHERRegions on AWS, Google Cloud and Azure across US, EU, UK, APAC, Brazil, Canada and UAE Free tier per active metric series, per GB ingested, per user, per session, per host SOC 2 Type II, ISO 27001, GDPR DPA
Splunk (Cisco)Splunk Observability Cloudsource · 8 Nov 2026 YesYesYesYesYesYes Native OTLP ingest SaaS US, EU, UK, APACRealms in US (3), EU (Dublin, Frankfurt), UK (London), Sydney, Tokyo, Singapore Not stated per host SOC 2 Type II, HIPAA
ElasticElastic Observabilitysource · 8 Nov 2026 YesYesYesNot statedNot statedYes Native OTLP ingest SaaS, Self-hostedopen source (AGPL-3.0 / SSPL 1.0 / Elastic License 2.0 (source); releases under the Elastic License) Not statedCloud regions not checked for this matrix Trial per GB ingested, per GB retained SOC 2 (type not stated), ISO 27001, HIPAA, FedRAMP, PCI DSS
SigNozSigNoz Cloudsource · 8 Nov 2026 YesYesYesYesNot statedNot stated Native OTLP ingest SaaS, BYOC, Self-hosted US, EU, APACData centres in US, EU and India Not stated per GB ingested, per event Not confirmed in sources checked
SigNozSigNoz Community Edition (self-hosted)source · 8 Nov 2026 YesYesYesYesNot statedNot stated Native OTLP ingest Self-hostedopen source (MIT (Expat), except ee/ directories under a separate licence) Where you host it Free (self-hosted) free (open source, you run it) Not confirmed in sources checked
HoneycombHoneycombsource · 8 Nov 2026 PartialYesYesNot statedNot statedPartial Native OTLP ingest SaaS Not statedNot stated on the pages checked Free tier per event Not confirmed in sources checked

yes · partial or add-on · no · – not stated in the sources we checked. Compliance lists only reports named on the vendor's own trust or security page; a blank does not mean the vendor lacks them. From each vendor's own documentation on the date shown; prices are not shown here, only how each product bills.

There is no single best APM tool. What matters is how you want to instrument (a vendor agent or OpenTelemetry), where the data may live, and which billing model fits how your infrastructure grows. These nine products cover the realistic shortlist.

Shortlist by situation

  • You want one vendor for everything and accept a vendor agent. Datadog, New Relic and Dynatrace each cover APM, infrastructure, logs, real user monitoring and synthetics in one product. They differ mainly in billing:
    • Datadog: per host for APM and infrastructure, plus per-GB and per-event charges for logs.
    • New Relic: per GB ingested and per user, or compute-based.
    • Dynatrace: memory-GiB-hours for full-stack monitoring.
  • You want to standardise on OpenTelemetry and keep the option to switch. Grafana Cloud, SigNoz and Honeycomb are built around OTel. Splunk Observability Cloud also lists native OTLP support.
  • You must self-host. SigNoz Community Edition (MIT-licensed core) and the AGPL-licensed Grafana stack run on your own infrastructure, which means you also operate storage, scaling and upgrades. Elastic can be self-managed too.
  • You want predictable spend per server. Splunk Observability Cloud prices its tiers per host per month, billed annually; APM with profiling is in the App & Infra tier.
  • Your main question is “why is this request slow?” Honeycomb’s model is built on high-cardinality events and traces, with metrics capped separately, rather than host-based APM.

What to compare before you trial

  1. Instrumentation path. A vendor agent gives the deepest auto-instrumentation in that vendor. OpenTelemetry makes the data portable, but check which vendor features depend on the vendor agent.
  2. Data regions. Datadog, Grafana Cloud, New Relic, Splunk and SigNoz Cloud document EU options. Several vendors do not list their regions on the pages we checked.
  3. How the bill grows.
    • Per-host pricing rewards few large hosts.
    • Per-GB pricing rewards controlling what you send.
    • Per-event pricing rewards sampling.

The full matrix, including logs and error tracking, is in observability platforms compared. For head-to-head detail, see Datadog vs New Relic and Datadog vs Grafana Cloud.

Vendor facts

Datadog

Datadog

Full-stack SaaS observability: separately billed products for infrastructure, APM, logs, RUM, session replay, error tracking, synthetics and profiling.

OpenTelemetry
Native OTLP ingest: OTLP via the Datadog Agent (with embedded DDOT Collector), via your own OTel Collector, or direct agentless OTLP ingestion
Runs as
saas
Free option
Free tier: Free tier for infrastructure: up to 5 hosts with 1-day metric retention
Bills
per host, per GB ingested, per event, per session, per test run (Infrastructure, APM and profiler per host; logs per GB ingested plus per million indexed events; replay per 1,000 sessions; errors by volume)
Compliance
SOC 2 Type II, ISO 27001, HIPAA, FedRAMP, PCI DSS

Verified 8 Nov 2026 · Datadog pricing, Datadog docs: OpenTelemetry, Datadog docs: sites, Datadog Trust Center, OpenTelemetry vendor registry (frozen list)

New Relic

New Relic

Full-stack SaaS observability priced mainly on data ingest and user type, with APM, infrastructure, logs, browser and mobile monitoring, session replay, errors inbox and synthetics.

OpenTelemetry
Native OTLP ingest: Native OTLP endpoints in US and EU (otlp.nr-data.net, otlp.eu01.nr-data.net)
Runs as
saas
Free option
Free tier: Perpetual free tier with 100 GB/month data ingest, one full platform user, unlimited basic users
Bills
per GB ingested, per user, usage credits (Per GB ingested plus per user type; alternative compute-based model (CCUs) without user fees)
Compliance
Not confirmed in sources checked

Verified 8 Nov 2026 · New Relic pricing, New Relic docs: OTLP endpoint, New Relic docs: data centers, OpenTelemetry vendor registry (frozen list)

Dynatrace

Dynatrace

Full-stack platform built around OneAgent auto-instrumentation, priced by consumption (host and memory-GiB-hours, GiB for logs and traces, sessions for RUM), as SaaS or the self-installed Dynatrace Managed.

OpenTelemetry
Native OTLP ingest: OTLP/HTTP traces, metrics and logs sent directly to the Dynatrace API or through an OpenTelemetry Collector
Runs as
saas, self-hosted
Free option
Trial: 15-day free trial
Bills
per host, per GB ingested, per GB retained, per session, usage credits (Full-stack monitoring per memory-GiB-hour; infrastructure per host-hour; logs and traces per GiB ingested, retained and scanned)
Compliance
SOC 2 Type II, ISO 27001, HIPAA, FedRAMP

Verified 9 Nov 2026 · Dynatrace pricing, Dynatrace docs: OneAgent, Dynatrace docs: OpenTelemetry, Dynatrace docs: Managed, Dynatrace Trust Center, OpenTelemetry vendor registry (frozen list)

Grafana Labs

Grafana Cloud

Managed LGTM stack (metrics, logs, traces, profiles) with Application Observability (APM), Frontend Observability (RUM), Kubernetes monitoring and synthetics.

OpenTelemetry
Native OTLP ingest: Listed with native OTLP support in the OpenTelemetry vendor registry
Runs as
saas
Free option
Free tier: "Always free": 10k active series, 50 GB each of logs, traces and profiles per month, 14-day retention, 3 active users
Bills
per active metric series, per GB ingested, per user, per session, per host (Usage-based above the free tier; APM and Kubernetes per host-hour)
Compliance
SOC 2 Type II, ISO 27001, GDPR DPA

Verified 8 Nov 2026 · Grafana Cloud pricing, Grafana Cloud docs: regional availability, Grafana Labs security, OpenTelemetry vendor registry (frozen list)

SigNoz

SigNoz Community Edition (self-hosted)

The open-source SigNoz you run yourself: APM, traces, logs, metrics, infrastructure and Kubernetes monitoring, without enterprise features.

OpenTelemetry
Native OTLP ingest: Built on OpenTelemetry; native OTLP
Runs as
self-hosted (open source, MIT (Expat), except ee/ directories under a separate licence)
Free option
Free (self-hosted): No licence fee; you operate infrastructure, storage, scaling, upgrades and backups
Bills
free (open source, you run it)
Compliance
Not confirmed in sources checked

Verified 8 Nov 2026 · SigNoz pricing (Community Edition), GitHub: SigNoz/signoz (LICENSE)

Sources

  1. Datadog pricing, checked 8 Nov 2026
  2. Datadog docs: OpenTelemetry, checked 8 Nov 2026
  3. New Relic pricing, checked 8 Nov 2026
  4. New Relic docs: OTLP, checked 8 Nov 2026
  5. Dynatrace pricing, checked 8 Nov 2026
  6. Grafana Cloud pricing, checked 8 Nov 2026
  7. Splunk Observability pricing, checked 8 Nov 2026
  8. Elastic Serverless Observability pricing, checked 8 Nov 2026
  9. SigNoz pricing, checked 8 Nov 2026
  10. Honeycomb pricing, checked 8 Nov 2026
  11. OpenTelemetry vendor registry, checked 8 Nov 2026

Matrix rows carry their own source and verification date. This page is revisited when: an APM product changes instrumentation or OpenTelemetry support, deployment options, free tier, pricing model or data regions

More observability decisions