Service and API discovery without instrumenting

You can now discover and monitor all your services and their HTTP APIs without touching a single line of code and without doing any instrumentation. If you’ve ever instrumented or tried to instrument your applications, even with OpenTelemetry, you’ll know this is a big deal.

Your services expose HTTP APIs. These APIs are called by all kinds of actors: your frontend, your other services, your mobile application if you have one, by third parties, MCP Servers, AI agents, etc. Whether requests come from a browser, mobile app, another service, or an external client, you need to know which endpoints are being called, their performance, health, and status. API Radar shows where requests are failing or slowing down and which services and which endpoints are handling them.

The new Service Monitoring

Both the Network Map and the new API monitoring (aka API Radar) live under Service Monitoring. The API Radar automatically discovers all your services and all their endpoints from live, observed traffic. It shows request volume, response status codes, error rates, and latency without code instrumentation or configuring each endpoint separately.

How and when does this help you?

The new API monitoring is very precise. When an API Radar alert reports high latency for your payment service, it identifies not only the affected service, but also the specific payment endpoint that’s affected, and it points you directly to its performance, health, and status details. From there, you can review request volume, errors, and response times to understand when the slowdown began and whether it coincides with increased traffic or failing requests.

Additionally, if you have also instrumented the affected application with OpenTelemetry, from here you can drill into the application’s traces and logs to get to the very root of the problem. Traces and logs will further tell you if the problem lies in a slow or broken database query or some such. You can also compare other endpoints in the same service to see whether the problem is isolated to payments or affects checkout more broadly.

Next, open Network Map to assess the potential blast radius. This tells you how far the problem could spread through your system. See which services call the payment service and follow their dependencies. A slowdown in a shared payment service could affect checkout, subscription renewals, and order processing, rather than a single operation.

Those dependencies help you prioritize the investigation. A problem affecting one background task has a different urgency from one affecting several customer-facing workflows. Follow the dependency chain to identify where the impact might extend, then check the connected services for corresponding errors or delays.

Thresholds & Alerts

Set warning and critical thresholds for API response error rate, p95, and p99 latency to highlight endpoints that need attention. These thresholds control health indicators and quick filters across API Radar, while alert rules let you configure when to receive notifications.

The built-in alert rules notify you about elevated service and endpoint error rates, upcoming SSL certificate expiry, and missing Authorization HTTP headers. Each triggered alert identifies the specific service and endpoint, with a direct link to investigate it. Example p95 and p99 latency rules are provided, but they are disabled by default so you can set thresholds appropriate for your services before enabling them. See Set latency thresholds for your services for details.

API Radar also provides a certificate inventory with issuers and expiry dates.

Excluding and ignoring endpoints

You can mark endpoints as internal to exclude them from selected alert rules. You can also mark specific endpoints as ignored. The ignored endpoints are removed from main reports and alerts. For instance, you might want to ignore health-check endpoints such as /health and /ready.

API Radar requires Sematext Agent 4.6.0 or later and, together with Network Map, is enabled when you enable Service Monitoring.

See the API Radar documentation for more.

Start Free Trial

---