Fix: backend port mapping #66
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
|
@@ -18,25 +18,25 @@ docker compose up | |||||||||
| ## Frontend | ||||||||||
| Start the development server on `http://localhost:3000`: | ||||||||||
| ```bash | ||||||||||
| pnpm --dir=frontend dev | ||||||||||
| ``` | ||||||||||
|
|
||||||||||
| Build the application for production: | ||||||||||
|
|
||||||||||
| ```bash | ||||||||||
| pnpm --dir=frontend generate | ||||||||||
| ``` | ||||||||||
|
|
||||||||||
| Locally preview production build: | ||||||||||
| ```bash | ||||||||||
| pnpm --dir=frontend preview | ||||||||||
| ```{% endraw %}{% if has_backend %}{% raw %} | ||||||||||
|
|
||||||||||
| ## Backend | ||||||||||
| Start the API server on port 4000: | ||||||||||
| ```bash | ||||||||||
| uv --directory=backend run uvicorn src.entrypoint:app --reload --host '::' --port 4000 | ||||||||||
|
There was a problem hiding this comment. The uvicorn command still uses hardcoded port 4000, which is inconsistent with the Docker configuration that now uses the deployed port variable. This should use the same port as configured in the Dockerfile.
Suggested change
|
||||||||||
| ```{% endraw %}{% endif %}{% raw %} | ||||||||||
|
Comment on lines
+36
to
+39
Contributor
There was a problem hiding this comment. Docs should reflect the dynamic backend port; avoid hardcoding 4000 The project now parameterizes the backend port, and the container listens on that port. The README still instructs to run on 4000, which can mislead when Apply this diff to keep the docs consistent with the templates: -Start the API server on port 4000:
+Start the API server on port {% endraw %}{{ backend_deployed_port_number }}{% raw %}:
```bash
-uv --directory=backend run uvicorn src.entrypoint:app --reload --host '::' --port 4000
+uv --directory=backend run uvicorn src.entrypoint:app --reload --host 0.0.0.0 --port {% endraw %}{{ backend_deployed_port_number }}{% raw %}In template/README.md.jinja around lines 36 to 39, the README hardcodes port |
||||||||||
|
|
||||||||||
| ## Updating from the template | ||||||||||
| This repository uses a copier template. To pull in the latest updates from the template, use the command: | ||||||||||
|
|
||||||||||
| Original file line number | Diff line number | Diff line change | ||||
|---|---|---|---|---|---|---|
|
|
@@ -18,7 +18,7 @@ RUN pnpm install --frozen-lockfile | |||||
| COPY . . | ||||||
|
|
||||||
| # build the Nuxt app (generates .output folder in Nuxt 3) | ||||||
| RUN pnpm list && pnpm generate | ||||||
|
There was a problem hiding this comment. The command change from 'pnpm run generate' to 'pnpm generate' may not work as expected. In most npm/pnpm projects, 'generate' is a script defined in package.json that should be invoked with 'pnpm run generate' or 'pnpm generate' only if it's a built-in pnpm command. Verify that 'generate' works without the 'run' prefix.
Suggested change
|
||||||
|
|
||||||
| # ------------ Stage 2: Production ------------ | ||||||
| FROM nginx:{% endraw %}{{ nginx_image_version }}{% raw %}-alpine{% endraw %}{{ alpine_image_version }}{% raw %} AS production | ||||||
|
|
@@ -30,10 +30,11 @@ COPY --from=builder /app/.output/public /usr/share/nginx/html | |||||
|
|
||||||
| # set sensible defaults (overridable at `docker run`) | ||||||
| ENV BACKEND_HOST=host.docker.internal \ | ||||||
| BACKEND_PORT={% endraw %}{{ backend_deployed_port_number }}{% raw %} \ | ||||||
| FRONTEND_PORT={% endraw %}{{ frontend_deployed_port_number }}{% raw %} | ||||||
|
|
||||||
| # Using port 80 caused problems running K8s on windows hosts, so using a higher port number | ||||||
| EXPOSE {% endraw %}{{ frontend_deployed_port_number }}{% raw %} | ||||||
|
|
||||||
| # Nginx in foreground | ||||||
| CMD ["sh", "-c", "envsubst '$BACKEND_HOST $BACKEND_PORT $FRONTEND_PORT' < /etc/nginx/conf.d/default.conf.template > /etc/nginx/conf.d/default.conf && exec nginx -g 'daemon off;'"]{% endraw %} | ||||||
|
There was a problem hiding this comment. The envsubst command with unvalidated environment variables could be vulnerable to injection attacks. Consider validating the environment variables or using a more secure templating approach. |
||||||
| Original file line number | Diff line number | Diff line change | ||||
|---|---|---|---|---|---|---|
| @@ -1,7 +1,7 @@ | ||||||
| {% raw %}FROM python:{% endraw %}{{ python_version }}{% raw %}-slim-{% endraw %}{{ debian_release_name }}{% raw %} | ||||||
|
|
||||||
| # docker build -t my-graphql-backend . | ||||||
| # docker run -p {% endraw %}{{ backend_deployed_port_number }}{% raw %}:{% endraw %}{{ backend_deployed_port_number }}{% raw %} my-graphql-backend | ||||||
|
|
||||||
| # tell uv to treat /usr/local as the project env | ||||||
| ENV UV_PROJECT_ENVIRONMENT=/usr/local | ||||||
|
|
@@ -21,7 +21,11 @@ COPY . . | |||||
|
|
||||||
| RUN uv pip list --format=columns --verbose | ||||||
|
|
||||||
| # set sensible defaults (overridable at `docker run`) | ||||||
| ENV API_PORT={% endraw %}{{ backend_deployed_port_number }}{% raw %} | ||||||
|
|
||||||
| # When deployed with network-mode=host on Rancher Desktop on Windows (for WSL-compatibility), there's no actual port mapping, so we need to have it running on the deployed port within the container itself | ||||||
| EXPOSE {% endraw %}{{ backend_deployed_port_number }}{% raw %} | ||||||
|
Comment on lines
+27
to
+28
Contributor
There was a problem hiding this comment. 💡 Verification agent 🧩 Analysis chainEXPOSE is documentation-only; ensure compose/k8s also wire containerPort to the same value EXPOSE helps tooling but doesn’t publish ports. Double-check docker-compose.yaml.jinja (and any Helm/k8s templates, if present) map {{ backend_deployed_port_number }} on both host and container, and pass API_PORT in environment. If already done in this PR, you’re good. 🏁 Script executed: #!/bin/bash
# Verify compose and templates map the same container port and set API_PORT
rg -n -C2 -P '(ports:|EXPOSE|API_PORT|backend_deployed_port_number)' --type=yaml --type=dockerfile --type=jinjaLength of output: 148 🏁 Script executed: #!/bin/bash
set -euo pipefail
# List all Jinja templates
echo "Listing all Jinja templates:"
fd --type f -e jinja || true
# Identify compose-related templates
echo -e "\nCompose-related templates:"
fd --type f -e jinja | grep -i 'compose' || true
# Verify backend_deployed_port_number usage in compose templates
echo -e "\nVerify backend_deployed_port_number in compose templates:"
fd --type f -e jinja | grep -i 'compose' | xargs -r rg -n -C2 'backend_deployed_port_number' || true
# Verify ports: mapping in compose templates
echo -e "\nVerify ports: in compose templates:"
fd --type f -e jinja | grep -i 'compose' | xargs -r rg -n -C2 'ports:' || true
# Verify API_PORT environment variable in compose templates
echo -e "\nVerify API_PORT in compose templates:"
fd --type f -e jinja | grep -i 'compose' | xargs -r rg -n -C2 'API_PORT' || true
# Identify K8s/Helm-related templates
echo -e "\nK8s/Helm-related templates:"
fd --type f -e jinja | grep -Ei 'helm|k8s|deployment|service' || true
# Verify containerPort mapping in K8s templates
echo -e "\nVerify containerPort in K8s templates:"
fd --type f -e jinja | grep -Ei 'helm|k8s|deployment|service' | xargs -r rg -n -C2 'containerPort' || true
# Verify backend_deployed_port_number in K8s templates
echo -e "\nVerify backend_deployed_port_number in K8s templates:"
fd --type f -e jinja | grep -Ei 'helm|k8s|deployment|service' | xargs -r rg -n -C2 'backend_deployed_port_number' || trueLength of output: 5659 🏁 Script executed: #!/bin/bash
set -euo pipefail
# Find all docker-compose Jinja templates
compose_files=$(fd --type f -e jinja | grep -i 'docker-compose.yaml')
# Print the first 200 lines of each to inspect ports and environment settings
for f in $compose_files; do
echo "===== Inspecting $f ====="
sed -n '1,200p' "$f" || true
echo ""
doneLength of output: 3221 EXPOSE is documentation-only; compose template needs to wire up port and API_PORT The Dockerfile’s • In ports:
- "{{ backend_deployed_port_number }}:4000"
environment:
API_PORT: "{{ backend_deployed_port_number }}"• If you introduce Kubernetes/Helm manifests for the backend, ensure they include: containerPort: {{ backend_deployed_port_number }}
env:
- name: API_PORT
value: "{{ backend_deployed_port_number }}"🤖 Prompt for AI Agents |
||||||
|
|
||||||
| # By default, run the entrypoint to serve the GraphQL app | ||||||
| CMD ["sh", "-c", "python src/entrypoint.py --host 0.0.0.0 --port $API_PORT"]{% endraw %} | ||||||
|
There was a problem hiding this comment. Using shell form with variable expansion can be vulnerable to injection attacks. Consider using exec form with environment variable expansion in the Python script instead, or validate the API_PORT environment variable.
Suggested change
|
||||||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The documentation still references port 4000 specifically, but the backend now runs on the configured deployed port. Consider updating this to reference the variable port or remove the specific port number.