500 errors

techhadoop

Best practice for 500 errors

  • Enable error.log (debug) in Kong.
  • Use file-log or http-log plugin to capture structured traffic.
  • Use correlation-id plugin to align Kong + upstream logs.
  • Always check upstream service logs — a 500 originates there, not in Kong.

Kong automatically adds X-Kong-Request-ID (if enabled).

  • Add plugin: correlation-id
  • Configure a header (e.g. X-Request-ID) so you can trace across Kong logs, upstream app logs, and client logs.

Leave a comment