[bug]: while login via the app, backend hoppscotch shows as 503 service unavailable
mrfzy00 opened this issue · comments
Is there an existing issue for this?
- I have searched the existing issues
Current behavior
The backend shows as error code of 503 service unavailable
How do i fix it? Do you have any solutions?
Steps to reproduce
- We're manage to create the hoppscotch with Helm charts, so we separate the services.
- We also have been state our .env to configmap on K8s, here's the code:
PORT: "3000"
DATABASE_URL: postgresql://user_hoppscotch:password@11.12.13.14:5432/tools_hoppscotch # or replace with your database URL
# Auth Tokens Config
JWT_SECRET: "secrets"
TOKEN_SALT_COMPLEXITY: "10"
MAGIC_LINK_TOKEN_VALIDITY: "3"
REFRESH_TOKEN_VALIDITY: "604800000" # Default validity is 7 days (604800000 ms) in ms
ACCESS_TOKEN_VALIDITY: "86400000" # Default validity is 1 day (86400000 ms) in ms
SESSION_SECRET: 'Secrets'
# Hoppscotch App Domain Config
REDIRECT_URL: "https://hoppscotch.xxx.xxx"
WHITELISTED_ORIGINS: "https://hoppscotch.xxx.xxx,https://hoppscotch-admin.xxx.xxx,https://hoppscotch-be.xxx.xxx"
# Google Auth Config
GOOGLE_CLIENT_ID: "*******"
GOOGLE_CLIENT_SECRET: "***********"
GOOGLE_CALLBACK_URL: "https://hoppscotch-be.xxxx.xxxx/v1/auth/google/callback"
GOOGLE_SCOPE: "email,profile"
# Rate Limit Config
RATE_LIMIT_TTL: "60" # In seconds
RATE_LIMIT_MAX: "100" # Max requests per IP
MAILER_SMTP_URL: "smtps://user@domain.com:pass@smtp.domain.com"
MAILER_ADDRESS_FROM: "from@example.com"
VITE_ALLOWED_AUTH_PROVIDERS: "GOOGLE"
VITE_BASE_URL: https://hoppscotch.xxx.xxx
VITE_SHORTCODE_BASE_URL: https://hoppscotch.xxx.xxx
VITE_ADMIN_URL: https://hoppscotch-admin.xxx.xxx
# Backend URLs
VITE_BACKEND_GQL_URL: https://hoppscotch-be.xxx.xxx/graphql
VITE_BACKEND_WS_URL: wss://hoppscotch-be.xxx.xxx/graphql
VITE_BACKEND_API_URL: https://hoppscotch-be.xxx.xxx/v1
# Terms Of Service And Privacy Policy Links (Optional)
VITE_APP_TOS_LINK: https://docs.hoppscotch.io/terms
VITE_APP_PRIVACY_POLICY_LINK: https://docs.hoppscotch.io/privacy
ENABLE_SUBPATH_BASED_ACCESS: "false"
- After that, it was run as we expected, but only accessible is frontend, admin shows blank pages, and backend was stated above.
Environment
Production
Version
Self-hosted
hi @balub , i'm using the nginx ingress, not using nginx here's the config:
ingress:
enabled: true
className: "nginx"
annotations:
#kubernetes.io/ingress.class: nginx
cert-manager.io/cluster-issuer: letsencrypt-staging
# kubernetes.io/tls-acme: "true"
hosts:
- host: hoppscotch-be.xxx.xxx
paths:
- path: /
pathType: ImplementationSpecific
tls:
- secretName: hoppscotch-be-tls
hosts:
- hoppscotch-be.xxx.xxx
service:
port: 3000
@mrfzy00 could you check if the backend service is up and active, maybe calling the ping route? When we spin up the backend for the first time we populate the DB with config props and terminate it, If it's docker we use the --restart unless-stopped
flag to restart the container when we do this, maybe in your case it didn't restart automatically. Could you just check this once?
This explains why you can only see the Hoppscotch app and not the admin app and the Backend because the Hoppscotch app is cached with a service worker so you can use it offline. But can you log in and do other things related to workspaces in it though?
hi @balub , sorry of late updating, my team just resolved the issues, it has been related to our template deployment on helm charts:
apiVersion: v1
kind: Service
metadata:
name: {{ include "hoppscotch-app.fullname" . }}
labels:
{{- include "hoppscotch-app.labels" . | nindent 4 }}
spec:
type: ClusterIP
ports:
- port: 3000
targetPort: http-app
protocol: TCP
name: http-app
selector:
{{- include "hoppscotch-app.selectorLabels" . | nindent 4 }}
while accessing the services, it seems the service wont recognized, but with that code above, we resolved this issue. i will close the issue.