render-infrastructure

name	render-infrastructure
description	Render deployment infrastructure reference. Use when accessing production/staging services, checking logs, SSH access, or deployment operations.

Render Infrastructure

Reference for accessing and managing Render-hosted services.

Service IDs

Environment	Service	ID
Production	Backend	<your-prod-backend-service-id>
Production	Frontend	<your-prod-frontend-service-id>
Staging	Backend	<your-staging-backend-service-id>
Staging	Frontend	<your-staging-frontend-service-id>

Render Account: <your-account-id>

---

Database Access

Database IDs

Environment	Database	ID
Production	backend-postgres	<your-prod-database-id>
Staging	staging-postgres	<your-staging-database-id>

Getting Connection Credentials

The Render CLI doesn't expose database credentials directly. Use SSH to retrieve them:

# Get production DATABASE_URL
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com 'echo $DATABASE_URL'

# Get staging DATABASE_URL
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-staging-backend-service-id>@ssh.frankfurt.render.com 'echo $DATABASE_URL'

The DATABASE_URL format is:

postgresql://<user>:<password>@<host>/<database>

• Internal host (for Render services): dpg-xxxxx-a (database ID)
• External host (for local access): dpg-xxxxx-a.frankfurt-postgres.render.com

Querying the Database

Option 1: Via SSH + Prisma (Recommended)

Use Prisma client through SSH for type-safe queries:

# Simple query
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.company.findMany({take:5}).then(r=>console.log(JSON.stringify(r,null,2))).finally(()=>p.\$disconnect())'"

# Count records
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.user.count().then(r=>console.log(\"Users:\",r)).finally(()=>p.\$disconnect())'"

# Raw SQL query
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.\$queryRaw\`SELECT tablename FROM pg_tables WHERE schemaname=\\\"public\\\"\`.then(r=>console.log(r)).finally(()=>p.\$disconnect())'"

Option 2: Via Render CLI psql (Interactive)

For interactive SQL sessions (requires TTY - won't work in automated scripts):

# Opens interactive psql session
render psql <your-prod-database-id>

Common Database Queries via SSH

List all tables

ssh ... "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.\$queryRaw\`SELECT tablename FROM pg_tables WHERE schemaname=\\\"public\\\" ORDER BY tablename\`.then(r=>console.log(r.map(t=>t.tablename).join(\"\\n\"))).finally(()=>p.\$disconnect())'"

Check record counts

ssh ... "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();Promise.all([p.user.count(),p.company.count(),p.engagement.count()]).then(([u,c,e])=>console.log({users:u,companies:c,engagements:e})).finally(()=>p.\$disconnect())'"

Find specific records

ssh ... "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.user.findFirst({where:{email:\"[e-mail]\"}}).then(r=>console.log(JSON.stringify(r,null,2))).finally(()=>p.\$disconnect())'"

Update records (USE WITH CAUTION)

ssh ... "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.enrichment.updateMany({where:{status:\"FAILED\"},data:{status:\"PENDING\"}}).then(r=>console.log(\"Updated:\",r.count)).finally(()=>p.\$disconnect())'"

Database Query Best Practices

1. Always use -o ConnectTimeout=30 - SSH connections can be slow
2. Use Prisma over raw SQL - Type safety and schema awareness
3. Always call $disconnect() - Prevents connection leaks
4. Use JSON.stringify for output - Makes results readable
5. Test queries on staging first - Before running on production
6. Ask user before updates - Never modify production data without permission

---

MCP Tools (Preferred)

When the Render MCP server is available, prefer MCP tools over CLI for programmatic access. MCP tools provide structured responses and better error handling.

Available MCP Tools

mcp__render__list_services      - List all services in the workspace
mcp__render__get_service        - Get details of a specific service
mcp__render__list_deploys       - List deployment history for a service
mcp__render__get_deploy         - Get details of a specific deployment
mcp__render__create_deploy      - Trigger a new deployment
mcp__render__cancel_deploy      - Cancel an in-progress deployment
mcp__render__list_jobs          - List jobs for a service
mcp__render__run_job            - Trigger a job run
mcp__render__list_logs          - Fetch logs with filters
mcp__render__list_env_vars      - List environment variables
mcp__render__update_env_var     - Update an environment variable
mcp__render__restart_service    - Restart a running service
mcp__render__scale_service      - Scale service instances
mcp__render__get_bandwidth      - Get bandwidth metrics
mcp__render__select_workspace   - Set workspace context

MCP Usage Examples

Check Service Health

mcp__render__get_service with:
  serviceId: "<your-prod-backend-service-id>"

List Recent Deployments

mcp__render__list_deploys with:
  serviceId: "<your-prod-backend-service-id>"
  limit: 5

Fetch Error Logs

mcp__render__list_logs with:
  serviceId: "<your-prod-backend-service-id>"
  type: "app"
  level: "error"
  direction: "backward"
  limit: 100

Trigger Deployment

mcp__render__create_deploy with:
  serviceId: "<your-prod-backend-service-id>"
  clearCache: false

Get Environment Variables

mcp__render__list_env_vars with:
  serviceId: "<your-prod-backend-service-id>"

---

Render CLI

The Render CLI (render) is installed locally. Always use -o json for scripting.

Correct Command Syntax

# List all services
render services list -o json

# View deploys - NOTE: serviceId is positional, not a flag
render deploys list <your-prod-backend-service-id> -o json

# View logs - NOTE: use -r flag for resources
render logs -r <your-prod-backend-service-id> -o json
render logs -r <your-prod-backend-service-id> --type build -o text  # Build logs

# Trigger manual deploy
render deploys create <your-prod-backend-service-id> -o json

# Cancel a deploy
render deploys cancel <deploy-id> -o json

Common CLI Patterns

Check Deploy Status

render deploys list <your-prod-backend-service-id> -o json | head -30

Get Build Logs (for failed deploys)

render logs -r <your-prod-backend-service-id> --type build -o text | tail -100

Wait for Deploy to Complete

# Check status, wait, check again
render deploys list <your-prod-backend-service-id> -o json | jq '.[0].status'
sleep 60
render deploys list <your-prod-backend-service-id> -o json | jq '.[0].status'

---

SSH Access

Verified Working

All instances tested and confirmed working:

Service	ID	SSH Status	Working Dir
Production Backend	<your-prod-backend-service-id>	OK	/opt/render/project/src/backend
Production Frontend	<your-prod-frontend-service-id>	OK	/opt/render/project/src/frontend
Staging Backend	<your-staging-backend-service-id>	OK	/opt/render/project/src/backend
Staging Frontend	<your-staging-frontend-service-id>	OK	/opt/render/project/src/frontend

Connection Pattern

# Always use these flags to avoid host key issues
# Use ConnectTimeout to handle slow connections (staging can take 20+ seconds)
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <service-id>@ssh.frankfurt.render.com "command"

# Examples:
# Staging Backend
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-staging-backend-service-id>@ssh.frankfurt.render.com "whoami && pwd"

# Production Backend (ask user before accessing)
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ConnectTimeout=30 \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com "whoami && pwd"

Why SSH Might Fail

Issue	Cause	Solution
Timeout	Connection too short	Use -o ConnectTimeout=30
Hang on connect	Service cold start / sleeping	Wait, or trigger activity first
Memory error	SSH needs ~2-5MB from service pool	Check service isn't memory-constrained
TTY required	Using render ssh command	Use direct ssh ... "command" instead

Environment Details

Once connected:

• Working directory: /opt/render/project/src/backend (for backend services)
• Project root: /opt/render/project/src/
• Environment vars: Available ($DATABASE_URL, etc.)
• Runtime: Node.js v22.22.0, npm 10.9.4
• Prisma: Client installed and working
• Project type: ES modules ("type": "module" in package.json)

CRITICAL: SSH Command Execution Patterns

The Render SSH shell has specific requirements. Use these patterns:

Pattern 1: Simple Commands (WORKS)

ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "whoami && pwd && ls -la"

Pattern 2: Node One-Liners with Single Quotes (WORKS)

ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'console.log(\"Hello from Render\")'"

Pattern 3: Prisma Queries (WORKS)

# Query the database
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.company.findMany({take:5}).then(r=>console.log(JSON.stringify(r,null,2))).finally(()=>p.\$disconnect())'"

Pattern 4: Update Data (WORKS)

# Update records
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.company.updateMany({where:{name:\"Example\"},data:{type:\"CLIENT\"}}).then(r=>console.log(\"Updated:\",r.count)).finally(()=>p.\$disconnect())'"

Patterns That DON'T Work

# DON'T: Interactive shell commands
render ssh <your-prod-backend-service-id>  # Requires TTY

# DON'T: Heredoc input
ssh server << 'EOF'
commands
EOF

# DON'T: bash -c with complex quoting
ssh server 'bash -c "complex command"'

# DON'T: ES module syntax (project uses ES modules)
# Instead, use require() in node -e one-liners

# DON'T: Run scripts from /tmp without copying to project dir first
# Prisma client won't be found

File Transfer (SCP)

# Upload a file
scp -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  /local/file.js \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com:/tmp/file.js

# Then copy to project and run (for CommonJS scripts)
ssh ... "cp /tmp/file.cjs /opt/render/project/src/backend/ && \
  cd /opt/render/project/src/backend && \
  node file.cjs && \
  rm file.cjs"

Note: Scripts must use .cjs extension or CommonJS require() syntax because the project has "type": "module" in package.json.

---

Common Operations

Check Deployment Status

# Using CLI
render deploys list <your-prod-backend-service-id> -o json | head -30

# Using MCP (preferred)
mcp__render__list_deploys with serviceId, limit: 5

Debug Failed Deployment

# Get build logs
render logs -r <your-prod-backend-service-id> --type build -o text | tail -100

# Check for migration errors
render logs -r <your-prod-backend-service-id> --type build -o text | grep -i "error\|failed"

Resolve Failed Migration

# SSH in and mark migration as rolled back
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "cd /opt/render/project/src/backend && npx prisma migrate resolve --rolled-back <migration_name>"

Query Production Database

ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.\$queryRaw\`SELECT COUNT(*) FROM \"Company\"\`.then(r=>console.log(r)).finally(()=>p.\$disconnect())'"

Update Production Data

# Example: Update a company type
ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null \
  <your-prod-backend-service-id>@ssh.frankfurt.render.com \
  "node -e 'const{PrismaClient}=require(\"@prisma/client\");const p=new PrismaClient();p.company.updateMany({where:{name:\"Company Name\"},data:{type:\"NEW_TYPE\"}}).then(r=>console.log(\"Updated:\",r.count)).finally(()=>p.\$disconnect())'"

---

Production Logs (BetterStack)

For log analysis, use the betterstack-logs skill which provides:

• Real-time log queries via ClickHouse SQL
• Error pattern analysis
• HTTP request/response analysis
• Database connection monitoring

Quick log check:

1. Create connection: mcp__betterstack__telemetry_create_cloud_connection_tool
   team_id: <your-team-id>, source_id: <your-source-id>

2. Query errors: mcp__betterstack__telemetry_query with SQL like:
   SELECT dt, JSONExtract(raw, 'message', 'Nullable(String)') AS message
   FROM remote(<your-table>_render_production_5_logs)
   WHERE JSONExtract(raw, 'level', 'Nullable(String)') = 'error'
   ORDER BY dt DESC LIMIT 20

Log source identifiers:

Hostname	Environment
<your-prod-backend-hostname>	Production backend
<your-staging-backend-hostname>	Staging backend

See betterstack-logs skill for full query patterns.

---

Related Commands

• /debug-production - Analyze production runtime errors
• /debug-production-deploy - Debug failed deployments
• /env-diff - Compare local vs production env vars

Related Skills

• betterstack-logs - Query production logs via BetterStack MCP
• local-debugging - Debug local development issues

Documentation

• https://render.com/docs/cli
• https://render.com/docs/ssh
• https://render.com/docs/mcp-server

---

Quick Reference Card

Task	Tool	Command/Pattern
List deploys	MCP	mcp__render__list_deploys
List deploys	CLI	render deploys list <service-id> -o json
Get logs	MCP	mcp__render__list_logs
Get logs	CLI	render logs -r <service-id> -o json
Build logs	CLI	render logs -r <service-id> --type build -o text
SSH query	SSH	ssh ... "node -e 'prisma one-liner'"
Trigger deploy	MCP	mcp__render__create_deploy
Trigger deploy	CLI	render deploys create <service-id>
Get DB credentials	SSH	ssh ... 'echo $DATABASE_URL'
Query DB	SSH	ssh ... "node -e 'PrismaClient query'"
Interactive psql	CLI	render psql <database-id>