4

I am using traefik version 1.7.7 as an ingress controller in a kubernetes cluster version 1.11.5.

Each time I perform a rolling upgrade of a Deployment of one of my backend, Traefik generates a 502 error (mostly only one) with the following error when I set Traefik in debug mode.

{"level":"debug","msg":"'502 Bad Gateway' caused by: read tcp 172.21.51.209:56624-\u003e172.21.63.132:3000: read: connection reset by peer","time":"2019-01-17T14:32:37Z"}
{"level":"debug","msg":"vulcand/oxy/forward/http: Round trip: http://172.21.63.132:3000, code: 502, Length: 11, duration: 623.180762ms","time":"2019-01-17T14:32:37Z"}

Is there anything I can do to avoid those 502?

Alex Strate
  • 651
  • 7
  • 10
  • ```readinessProbe:exec: command: ["./checks.sh", "/tmp/gunicorn.pid"] initialDelaySeconds: 10 periodSeconds: 10 successThreshold: 3 failureThreshold: 2 timeoutSeconds: 5``` – Alex Strate Jan 18 '19 at 09:11

0 Answers0