troubleshooting tips

docs.json

@@ -200,6 +200,7 @@
               {
                 "group": "Troubleshooting",
                 "pages": [
+                  "references/troubleshooting/jupyter-labs-not-starting",
                   "references/troubleshooting/jupyterlab-checkpoints-folder",
                   "references/troubleshooting/leaked-api-keys",
                   "references/troubleshooting/storage-full",

references/troubleshooting/jupyter-labs-not-starting.mdx

@@ -0,0 +1,20 @@
+---
+title: "Port apps not starting"
+sidebarTitle: "Port apps not starting"
+---
+
+You might encounter an issue where you're unable to launch JupyterLab on a port or see a bad gateway connection for an application on one of your ports in your template. To troubleshoot this issue, you'll need to determine whether it's a Runpod issue or a template issue.
+
+## Test with an official Runpod PyTorch template
+
+Using an official Runpod PyTorch template helps establish a baseline for expected behavior. Custom templates may have configuration issues that prevent JupyterLab from starting properly. By testing with the official Runpod template, you can determine whether the issue is with Runpod itself or with your custom template configuration.
+
+![Official Runpod PyTorch template](/images/official-runpod-pytorch.png)
+
+## Check the Pod logs
+
+The Pod logs typically show errors when a service fails to start or confirm when services have started successfully. Review the logs carefully for any error messages or warnings that might indicate why JupyterLab isn't starting. The image depending how big it is just might be downloading too.
+
+## Allow time for startup
+
+Some templates require significant startup time, especially popular templates like ComfyUI that have heavy dependencies. Depending on how the applications are configured to start, initialization can take a long time. Be patient and monitor the logs to ensure the startup process is progressing.