> ## Documentation Index
> Fetch the complete documentation index at: https://docs.incident.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Temporal

> Let Nexus answer questions about your Temporal workflows and task queues.

Connect Temporal and Nexus can answer questions about your workflows in [investigations](https://docs.incident.io/investigations/overview.md), [the agent](https://docs.incident.io/ai/at-incident.md), and the [remote MCP server](https://docs.incident.io/ai/remote-mcp.md):

* Which `Charge` workflows failed in the last hour, and why?
* Are any workflows stuck retrying?
* Did failures start before or after the 14:00 deploy?
* Are any workers polling the `orders` task queue?
* What's filling the `email` task queue's backlog?

## What we support

Nexus reads workflows, their event histories and pending activities, and task queues. It never starts, signals, cancels, or terminates a workflow.

It doesn't read payloads: workflow inputs and results, signal arguments, and memos stay in Temporal. It does read failure messages and search attributes.

Temporal Cloud limits each namespace's visibility calls, and your own tools share that limit. Nexus stays well below it.

This data source doesn't read metrics. If you export Temporal's SDK or Cloud metrics to a data source such as Prometheus or Datadog, connect that too, so Nexus can see trends such as task latency across all your workflows.

## Connecting Temporal

Add a telemetry data source in [Nexus](https://app.incident.io/~/nexus/telemetry), choose **Temporal**, enter a login and the namespaces to connect, and test the connection. Each namespace becomes its own data source.

Use an API key if you can, because you can limit it to read access. A client certificate can't be limited: it has full access to every namespace that trusts its CA.

### Temporal Cloud

#### API key

1. Create a [service account](https://docs.temporal.io/cloud/service-accounts) with **Read** on each namespace to connect.
2. Create an [API key](https://docs.temporal.io/cloud/api-keys) for it, and paste it in.
3. Choose **Find namespaces**, and pick the ones to connect.

#### Client certificate

1. Issue a client certificate from a CA your namespaces trust. See Temporal's [certificate guide](https://docs.temporal.io/cloud/certificates#issue-certificates).
2. Paste the PEM-encoded certificate and its private key.
3. Enter each namespace's full name, such as `payments.a2b3c`. Temporal Cloud can't list namespaces for a certificate.

### Self-hosted

1. If the frontend isn't reachable from the internet, choose **Private network** and attach a [proxy](https://docs.incident.io/integrations/proxy.md).
2. Enter the frontend address, such as `temporal-frontend.internal:7233`. The frontend must use TLS.
3. Paste a client certificate and its private key, issued by a CA the frontend trusts. See Temporal's [mTLS guide](https://docs.temporal.io/self-hosted-guide/security#encryption-in-transit-with-mtls). Add a CA certificate if the cluster's server certificate uses a private CA.
4. Choose **Find namespaces**, and pick the ones to connect.

## Troubleshooting

* **Temporal refused the API key.** A namespace that also accepts client certificates refuses API keys on its own endpoint. Turn on **These namespaces also accept client certificates**, and enter your region's endpoint, such as `europe-west3.gcp.api.temporal.io:7233`. Connect each region as a separate data source.
* **The API key can't read a namespace.** Give its service account **Read** on that namespace.
* **Temporal refused the client certificate.** Check that the namespace trusts the certificate's CA, and that the namespace's [certificate filters](https://docs.temporal.io/cloud/certificates) match the certificate.
* **No namespace by that name.** Use the full name, including the account ID, such as `payments.a2b3c`.
* **Couldn't reach Temporal.** Check the host and port. Temporal Cloud uses port 7233. For a cluster on a private network, attach a [proxy](https://docs.incident.io/integrations/proxy.md).

## Related

> [card] [How telemetry works](https://docs.incident.io/nexus/telemetry/how-it-works.md)
>
> How Nexus learns and queries your namespaces.

> [card] [Proxy](https://docs.incident.io/integrations/proxy.md)
>
> Reach a self-hosted cluster on a private network.
