RabbitMQ - support Kubernetes internal DNS routing

**Output of the info page **

    rabbitmq (5.18.1)
    -----------------
      - instance #0 [OK]
      - Collected 0 metrics, 0 events & 1 service check

and from logs:

2017-10-31 14:22:10 UTC | ERROR | dd.collector | checks.rabbitmq(rabbitmq.py:190) | Error executing check: Cannot open RabbitMQ API url: queues Invalid URL 'queues': No schema supplied. Perhaps you meant http://queues?
2017-10-31 14:22:10 UTC | ERROR | dd.collector | checks.rabbitmq(rabbitmq.py:193) | error while contacting rabbitmq (rabbitmq-management.default.svc.cluster.local:443/api/), setting aliveness to CRITICAL for vhosts: {}

Additional environment details (Operating System, Cloud provider, etc):

• CoreOS
• AWS

Steps to reproduce the issue:

1. Override datadog docker image to support RabbitMQ
2. Enter address following convention <your_service>.<your_namespace>:<port>
3. Start pod with datadog Daemon Set

Describe the results you received: I got the logs described above meaning that somewhere in the process URL is changed to empty string. That’s why I got message Invalid URL 'queues': No schema supplied. Perhaps you meant http://queues?

Describe the results you expected: Be able to route your traffic internally without the need to use external DNS

Additional information you deem important (e.g. issue happens only occasionally): curl https://rabbitmq-management.default results in read timeout. curl rabbitmq-management.default:443 results in proper response so we are able to connect from the machine.

If I add AWS ELB alias or IP I am able to get traffic.

I know it’s more of a feature request but I hope you will look into it.

Cheers, Sebastian

[sdomagala]

Hello @gmmeyer, sorry for the late reponse. It appears the link you provided works! I don’t know why http is served over 443 in Kubernetes, it didn’t cross my mind to check it 😕

Thanks for the help!

[gmmeyer]

Can you do the same for

curl -vvvvv http://rabbitmq-management.default:443