Cyprian Kiplangat — February 11, 2026

Hello. I am trying to follow the example in the quickstart But I get errors when I try to install playwright mcp These are the docker logs:

7 replies

joey (archestra team) — February 11, 2026

hey there! The KinD cluster appears to be working fine here — the issue is the Playwright MCP server pod inside KinD is timing out (30s x 3 retries). Playwright is a heavier server since it needs to download/install Chromium (~150MB) on first start, which can easily exceed the timeout in a resource-constrained environment. Can you run these commands to help debug? ```# Check if the MCP server pod is running/crashed docker exec archestra-quickstart kubectl get pods # Check the pod logs docker exec archestra-quickstart kubectl logs -l app=mcp-server --tail=100 # Check for events like image pull errors or OOM kills docker exec archestra-quickstart kubectl describe pods -l app=mcp-server``` (Replace `archestra-quickstart` with your container name if different)

Cyprian Kiplangat — February 11, 2026

it fails. "executable file not found in $PATH"

joey (archestra team) — February 11, 2026

actually, can you try these commands please? ```# Step 1: List MCP server containers (look for one with "playwright" in the POD column) docker exec archestra-mcp-control-plane crictl ps -a --name mcp-server # Step 2: Get logs for the container (replace CONTAINER_ID with the ID from step 1) docker exec archestra-mcp-control-plane crictl logs CONTAINER_ID``` this will show us whether the mcp server pod is crashing, stuck pulling dependencies, or something else.

Cyprian Kiplangat — February 11, 2026

`$ docker exec archestra-mcp-control-plane crictl logs f707e615684de [17:43:36]` Listening on localhost:8080 Put this in your client config: { "mcpServers": { "playwright": { "url": "localhost:8080/mcp" } } } For legacy SSE transport support, you can use the /sse endpoint instead. === its the same logs I am seeing in mcP registry

joey (archestra team) — February 11, 2026

the server is actually running fine — the logs show `Listening on localhost:8080`, which tells us the issue. It's binding to `localhost` (127.0.0.1) instead of `0.0.0.0`, so Archestra can't reach it from outside the pod. Can you share how you configured the server when installing it (command, args, transport type)? There are two likely scenarios: 1. Transport mismatch — If you left the transport as "stdio" (default), but the server is running in HTTP mode, Archestra is trying to communicate via stdin/stdout while the server is listening on HTTP. In this case, you'd either need to switch to streamable-http transport in the server config, or change the command so the server runs in stdio mode (the default `npx @playwright/mcp@latest` without `--port` runs in stdio mode). 2. Network binding — If you configured it with streamable-http transport, the server also needs to bind to `0.0.0.0` instead of `localhost` so the K8s Service can route traffic to it.

Dominik Broj (archestra team) — February 12, 2026

Hi @user, Playwright MCP needs this config to work properly. In upcoming release, this mcp server will be seeded on startup with correct config together with few other fixes around it. I have just tested Docker quickstart against latest `main` branch and it works ok. If you want to run Archestra with Docker quickstart mode, I think the easiest way to unblock you before we release new version is to build Docker image locally and then run it. Pull latest `main` branch and then: ```docker build -t archestra/platform:latest <path_to_archestra/platform_directory> docker run -p 9000:9000 -p 3000:3000 \ -e ARCHESTRA_QUICKSTART=true \ -v /var/run/docker.sock:/var/run/docker.sock \ -v archestra-postgres-data:/var/lib/postgresql/data \ -v archestra-app-data:/app/data \ archestra/platform ``` Then you will see in MCP Registry preseeded `microsoft__playwright-mcp` item, install it and you should be good to go. Let me know in case you encounter any issues with it

Cyprian Kiplangat — February 12, 2026

it now worked, thanks