Kong Prometheus Plugin

Tracks internal kong metrics and exposes a prometheus endpoint on the Admin API to be scraped by a Prometheus Collector.

Usage

$ curl -X POST http://localhost:8001/plugins -d "name=prometheus"

Now kong will track and monitor metrics across all apis and consumers.

You can view these metrics by going to:

http://localhost:8001/prometheus/metrics

Metrics tracked

• kong_http_requests_total - Number of HTTP requests
• kong_http_request_duration_seconds - HTTP Request Latency
• kong_http_request_size_bytes - HTTP Request Size Bytes
• kong_http_response_size_bytes - HTTP Response Size Bytes
• kong_upstream_http_request_wait_duration_seconds - Upstream HTTP Request Time to First Byte Latency
• kong_upstream_http_request_duration_seconds - Upstream HTTP Request Latency
• kong_connections_active - Total number of active connections
• kong_connections_reading - Total number of connections reading
• kong_connections_writing - Total number of connections writing
• kong_connections_waiting - Total number of connections waiting

Special Thanks

• To knyar for their nginx-lua-prometheus client which this was built around and uses to track metrics.

Todo

• Implement Prometheus endpoint
• Tests