# Install Postman Insights Agent as a sidecar

After you create an Insights project and select **Start with Kubernetes**, you can begin your Insights Agent installation as a [sidecar](https://kubernetes.io/docs/concepts/workloads/pods/sidecar-containers/). To reproduce errors on real user data, activate the Insights' Repro Mode.

<Info class="iconless-callout">
  **Estimated time:** 10-minute setup, 5-minute wait
</Info>

After onboarding a service, you'll see your endpoints and their insights. Then, you can leverage the Insights' [Repro Mode](#activate-repro-mode) to debug your failing endpoints. You can also set up [alerts](#set-up-alerts) to get notifications about your errors.

## Requirements

* Edit permissions for the Kubernetes deployment.
* Postman Workspace Admin or Editor role.
* **Linux environment.** The agent and the user’s service should be running in a Linux environment.
* **Insights project.** See [Get started with Postman Insights](/docs/insights/get-started/overview/) to learn more.

## Install the Insights Agent

To install the Insights Agent, do the following:

1. As shown on the Insights onboarding screen, run the following command on your command line:

   ```bash
   bash -c "$(curl -L https://releases.observability.postman.com/scripts/install-postman-insights-agent.sh)"
   ```

2. Navigate to the correct Kubernetes cluster and get the existing deployment configuration.

   ```bash
     kubectl get deployments -A
   ```

3. On the Insights onboarding screen, click **Add API key** to copy and paste it with the next command.

4. Use the `kube inject` command to set up the Postman Insights Agent in your Kubernetes environment and edit the deployment file. For more information about this command and its flags, see [kube inject](/docs/insights/reference/agent/kube-inject/).

   <Info class="iconless-callout">
     In the example command below, the angle brackets and the text between them are placeholders. The command you copy in the onboarding screen will contain your correct API key and project ID.
   </Info>

   ```bash
   kubectl get -n <namespace> deployment/<deployment> -o yaml \
   | POSTMAN_API_KEY=<add-your-api-key-here> postman-insights-agent kube inject --project <projectId> --repro-mode -s=true -f - \
   | kubectl apply -f -
   ```

   The following example shows the values the Insights Agent adds to the deployment.

   <img alt="Kubernetes deployment file" src="https://assets.postman.com/postman-docs/v11/insights-k8s-deployment-v11b.jpg" />

   The command creates a secret to hold your Postman key and updates the deployment with the Postman sidecar. The response looks something like this:

   ```shell
   secret/postman-agent-secrets created
   deployment.apps/cloud-api-deployment configured
   ```

5. Check the status of your Kubernetes container pods.

   ```bash
   kubectl get pods -n <your_namespace>
   ```

   The status response looks something like this:

   ```shell
   NAME READY STATUS RESTARTS AGE
   cloud-api-deployment-5db4fBc64-cknhc 2/2 Running 0 4h40m
   cloud-api-deployment-7f97856c7f-p5Lv7 2/2 Running 0 15h
   ```

6. Get the logs from the sidecar.

   ```bash
   kubectl logs -n <your_namespace> <your_pod> postman-insights-agent
   ```

7. Let the Insights Agent run for a minute and observe the HTTP and TLS traffic on your ports.

8. Return to Postman and observe an Insights Project populated for your services.

<Info class="iconless-callout">
  **Tip:** If you're not seeing your endpoints right away, traffic may be taking some time to collect. See the next section for more information.
</Info>

## Wait for the Insights Agent to collect traffic

After Insights detects traffic from the Insights Agent, you'll get automatically redirected to the **Overview** page. The Insights Agent needs 5-8 minutes to use AI to generate endpoints.

If you're only seeing health checks, your traffic may be affected. If you're not seeing what you expect after 10 minutes, see [Diagnose and troubleshoot Insights Agent errors](/docs/insights/troubleshoot/overview/).

## Activate Repro Mode

<Info class="iconless-callout">
  Only a Workspace Admin can enable the **Repro Mode**.
</Info>

To make API calls using real user data, you need to activate Repro Mode. Go to your Insights project and click [**Settings**](/docs/insights/reference/app/settings-tab/). Then, toggle on **Activate Repro Mode**.

<Info class="iconless-callout">
  **Default data redactions.** Postman Insights automatically redacts a set of sensitive values including authentication tokens. See the [full list](/docs/insights/data/redactions/). You can also add fields to redact ahead of turning on the feature.
</Info>

Learn more about [debugging with Repro Mode](/docs/insights/debug/).

## Set up alerts

Insights alerts notify you about your errors in real time, so you can act on them immediately. In your Insights project's **Alerts** tab, you can adjust error rate thresholds, find and view problematic endpoints, mute or unmute alerts globally for specific endpoints, and integrate your alerts with Slack. Learn more about [configuring and using alerts](/docs/insights/reference/app/alerts-tab/).