Trace the root cause of every failure

Checkly Traces connects OpenTelemetry spans to your synthetic checks, so a failed check opens straight to the trace that explains it. Your checks and traces live in one place, so the context you need to fix a failure is already there.

Start for free View docs

Trusted by engineering teams running `OpenTelemetry` in production

Distributed systems fail in cascades. Most tracing tools bury the cause.

When one service breaks, the failure spreads to everything downstream, and finding the origin means digging through logs across half a dozen tools. Checkly inverts the model. It records only the spans tied to a failed check, so the cause is already in front of you.

Traditional tracing

Sample everything, sort it out later

$$$$

4.1 TBingested / day

Checkly Traces

Only the spans tied to a failed check

3.1 GBingested / day

Traditional tracing captures everything, so the signal you need is buried in noise. Checkly captures only the spans tied to a failed check, the ones you actually need to debug.

Find the root cause, instantly

When a synthetic check fails, jump straight to the exact span where it broke. No dashboard-hopping, no guessing which service to blame.

Checkout

GET /cart

POST /checkout

SELECT orders

Cut MTTD and MTTR with Rocky AI

Rocky AI reads the correlated trace and hands you a plain-language root cause, so your team resolves incidents in minutes instead of hours.

Rocky AIroot cause · 0:04

Failure traces to a Clickhouse query returning HTTP 500, not the API layer.

A failed check, already correlated to its trace

Checkly links every check result, test session, and editor run to its OpenTelemetry trace. The moment a check fails, the trace is right there, and Rocky AI reads the spans to tell you what actually broke.

Check results, test sessions, and live editor runs are all traced end-to-end.
Failed checks auto-correlate to the exact spans that caused them.
Rocky AI interprets the trace into a plain-language root cause.
Drill from a red check to the failing span in a single click.

Root CauseRocky AI

The /api/checkout check failed with a 500, but the trace localizes the fault to a downstream Clickhouse query, SELECT orders, returning HTTP 500. An infrastructure failure in the analytics database, not your application code or the API gateway.

Clickhouse Status Page
Checkly Status Page

Evidence2

The check timed out asserting on the order confirmation. /api/checkout returned a 500 instead of the expected 201.
error
OpenTelemetry spans show a 500 on the SELECT orders query to Clickhouse (db.system=clickhouse), pinpointing the analytics database as the failing dependency.
otel spans

instrumentation.ts

1// instrumentation.ts

2import { NodeSDK } from '@opentelemetry/sdk-node'

3import { OTLPTraceExporter } from '@opentelemetry/exporter-trace-otlp-http'

5const sdk = new NodeSDK({

6 // Point your existing OTel setup at Checkly. No new agent.

7 traceExporter: new OTLPTraceExporter({

8 url: 'https://otel.eu-west-1.checklyhq.com/v1/traces',

9 headers: { authorization: `Bearer ${process.env.CHECKLY_KEY}` },

10 }),

11})

13sdk.start()

Don't take our word for it.

“Checkly Traces helped us resolve issues faster by showing exactly how long database calls are taking. This insight was crucial in pinpointing N+1 issues and optimizing caching. Having a vendor like Checkly levels up your team to the point where it starts to feel like an unfair advantage.”

James Hall

AWS Hero & Founder · Parallax

Export to the observability tools you already run

Checkly Traces speaks OTLP, so check-generated spans flow into whatever backend your team relies on. No proprietary agents, no lock-in.

OpenTelemetry

Native OTLP ingestion. Send spans from any OTel-instrumented service or Collector.

View docs

Grafana Tempo

Forward check-generated traces into Tempo alongside the rest of your telemetry.

View docs

Honeycomb

Stream spans to Honeycomb for high-cardinality querying and deep debugging.

View docs

Datadog

Pipe check traces into your existing Datadog APM views and dashboards.

View docs

Coralogix

Ship spans to Coralogix for log-and-trace analytics with no extra agents.

View docs

Any OTLP backend

If it speaks OTLP, Checkly can export to it. Your instrumentation stays portable.

View docs

Traditional tracing samples all of your traffic and bills you for the volume. Checkly captures only the spans tied to a failed synthetic check, so you get the exact context to debug a failure without the unrelated telemetry or its storage bill.

No. Checkly Traces is built on standard OpenTelemetry. Instrument with the OTel SDKs you already run, point an existing Collector at Checkly, and export the same traces to Grafana Tempo, Honeycomb, or Datadog. There is no vendor lock-in.

Every check result, test session, and live editor run is correlated to its OpenTelemetry trace. When a check fails, Checkly links it directly to the spans that caused it, so you go from a red check to the failing span in one click.

When a check fails, Rocky AI automatically retrieves and interprets the OTel trace to pinpoint the root cause. For example, an API check can fail with a 500 while the trace reveals the real culprit is a downstream Clickhouse query returning the error, and Rocky reports that in plain language.

Anything with OpenTelemetry instrumentation: Node.js, Python, Go, Java, .NET, Ruby, and more. If your services emit OTel spans, Checkly can correlate them to the checks that generated them.

Yes. Because Checkly only records the interactions it generates, trace volume drops sharply compared with sample-everything pipelines, which lowers storage and egress costs while keeping the data you actually need to resolve issues.

See the root cause on your next failed check

Connect OpenTelemetry to your synthetic checks and turn every failure into a trace you can act on, already correlated and ready for Rocky AI.

Start for free Book a demo

Trace the root cause of every failure

Trusted by engineering teams running `OpenTelemetry` in production