> For the complete documentation index, see [llms.txt](https://docs.controltheory.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces.md).

# Traces

Your services, how they call each other, their requests, errors and latency, and the traces worth opening

Dstl8 reads your OpenTelemetry traces to show how requests move through your services and where they fail or slow down. Instead of asking you to search through millions of spans, it gives you:

* **Service health:** requests per minute, error share and p95 latency for every service and operation, from every span you send.
* **A service map:** which service calls which, with the call rate, errors and latency on each connection.
* **Request paths:** the routes each kind of request takes through your services, compared with the period just before, and tagged when one is new, busier, quieter, failing more or slower.
* **Traces worth opening:** failing traces, unusually slow ones and a sample of normal ones, each with a waterfall of its spans and the log lines written during each span.

Möbius uses the same reads when it investigates an incident. It names the span that failed or slowed down, not just the service that reported the error, and can cite a trace as evidence beside log lines and metrics.

{% hint style="info" %}
Traces is in beta, and your plan must include it (**Organization Settings** → **General** lists it under **Features**). There are no alert rules to write and nothing to tune per service.
{% endhint %}

## How traces reach Dstl8

| Path                     | Use it when                                                                                                                               | Set up on                                                                                                                   |
| ------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------- |
| **OpenTelemetry source** | Your apps or an OpenTelemetry Collector can send OTLP/HTTP over the internet. Works from anywhere, including Kubernetes.                  | [OpenTelemetry (OTel)](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/sources/opentelemetry-otel.md) |
| **Docker agent**         | Your apps run on a Docker host with the Dstl8 agent (v1.4.0 or later). They send OTLP to the agent on the host (ports `4317` and `4318`). | [Docker](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/sources/docker.md)                           |

Traces need no extra switch on the source: once spans arrive, the Traces page fills in within a minute or two. See [Sending traces](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/sending-traces.md) for endpoints, SDK settings, service naming and troubleshooting.

{% hint style="warning" %}
On Kubernetes, send traces to an **OpenTelemetry source** (directly or through your own Collector). The Kubernetes agent's OTLP ports on each node take logs, and metrics when you turn that on, but not traces.
{% endhint %}

## What Dstl8 keeps

* **Requests, errors, latency and the service map** are worked out from every span you send, so the numbers are complete even though not every trace is stored.
* **Whole traces** are kept selectively: failing traces, unusually slow ones, traces of new request paths, and a regular sample of normal traffic, so there is always a healthy trace to compare a bad one with.
* **Health-check spans** are dropped as they arrive. They are noise for request rates and latency. See [Health checks](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/sending-traces.md#health-checks) for which ones.

Trace links don't last forever. Opening an old one shows **This trace is no longer kept**; the incident or note that cited it is still there.

## Where you see them

* **Explore → Traces:** each workspace's Traces page: summary tiles, the service map, a services table, request paths, and a scatter of the kept traces that opens a waterfall. See [Exploring traces](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/explore.md).
* **Service pages:** every service has a page with its signals, operations and calls in and out. Open one from **Service page** on the Traces page. See [Services](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/services.md).
* **Incidents:** when traces are part of an incident, its **Signals** section shows a card per operation, with **Open in Traces** and a button to open a failing or the slowest trace around the peak. See [Incidents](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/incidents.md).
* **Topology:** a workload's view shows its service's traces. See [Topology](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/topology.md).
* **Canvas:** if your organization has Canvas, **Open in Canvas** on the Traces page starts a canvas with requests, p95 latency and error rate by service, service calls, and traces. See [Canvas](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/canvas.md).
* **MCP:** your own assistant gets the same trace tools Möbius uses. See [Traces MCP Tools](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/mcp-tools.md).
* **Möbius chat:** the chat sidebar on the Traces page knows the range, the service you picked, the busiest services and the open trace anomalies, so you can ask "why is checkout slow?" without explaining where you are.

## Which workspace sees which traces

A workspace sees traces from the sources and streams assigned to it, scoped by service (`service.name`), and only when **Traces** is among its signals. A trace that crosses services shows in a workspace when any of its spans belongs to that workspace. See [Signals and Workspaces](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/signals.md) for how to assign a service and choose signals.

## Next

* [Exploring traces](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/explore.md): every part of the Traces page, and how to chase a slow request or an error
* [Sending traces](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/sending-traces.md): endpoints, SDK settings, service names, sampling and troubleshooting
* [Traces MCP Tools](https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces/mcp-tools.md): the trace tools and prompts to try in your own assistant


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.controltheory.com/controltheory-documentation/dstl8-docs/traces.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
