Deploy the Cachee agent on your EC2 instance. The agent monitors your ElastiCache, accelerates cache operations, and reports metrics to your Cachee dashboard.
How Agent-Based Monitoring Works
1
You install the Cachee agent on your EC2 instance (same VPC as ElastiCache)
2
Agent connects to your ElastiCache using the instance's own IAM role — no extra credentials needed
3
Agent reports metrics outbound via HTTPS to app.cachee.ai/api/metrics every 60 seconds
4
Your Cachee dashboard shows real-time hit rates, latency, and AI optimization recommendations
No inbound access from Cachee. The agent only makes outbound HTTPS calls. Your security group rules don't need to allow any traffic from Cachee's infrastructure.
Curl
Docker
Manual
SSH into your EC2 and run this single command. It installs Node.js (if needed), PM2, and the Cachee agent. ~90 seconds.
Inbound TCP port 3000 (or your configured port) from your application servers — so they can route cache requests through Cachee
2
Outbound HTTPS (443) to app.cachee.ai — for metrics reporting (usually already allowed by default SG rules)
3
Outbound TCP 6379 to your ElastiCache security group — for cache operations
EC2 and ElastiCache must be in the same VPC and region. Cross-VPC or cross-region adds 50-200ms latency and defeats the purpose of caching.
Troubleshooting
Agent won't start
Check logs: pm2 logs cachee-agent or journalctl -u cachee-agent -f
Can't connect to ElastiCache
Ensure the ElastiCache security group allows inbound TCP 6379 from your EC2's security group. Both must be in the same VPC.
Metrics not showing in dashboard
Verify outbound HTTPS is allowed: curl -v https://app.cachee.ai/api/health-check. Check that your API key is correct in /etc/cachee/agent.yaml.
Health check fails in wizard
The wizard probes http://YOUR_IP:3000/api/v1/health. Ensure your EC2 security group allows inbound TCP 3000 from 0.0.0.0/0 (or at minimum from Cachee's verification IPs).