Does not forward anything to Stackdriver Trace from GKE.

[yehudamakarov]

When I run the docker command

docker run -d -p 9411:9411 \ 
-e STORAGE_TYPE=stackdriver \
-e GOOGLE_APPLICATION_CREDENTIALS=/var/secret/google/key.json \
-e STACKDRIVER_PROJECT_ID=<ID> \
-v $HOME/Desktop/googleKeys:/var/secret/google openzipkin/zipkin-gcp

I successfully start the image and see logs for the zipkin server. But I only see a 404, when i visit the host/port of the server in the browser.

In the regular zipkin docker image I see the web UI. Acutally / reroutes to /zipkin which serves the UI.

I'm really running this as a deployment behind a service in GKE, and nothing is being forwarded to Trace. Looking for a way to debug.

[codefromthecrypt]

The stackdriver integration is write-only at the moment, which is why you don't see zipkin's UI. Are you intentionally interested in using Zipkin's UI instead of stackdriver's?

[yehudamakarov]

That's REALLY helpful to know. thank you. No i actually don't need UI at all. But no traces are making it into stackdriver, and I'm not sure how to start trouble shooting.

I followed https://kubernetes.github.io/ingress-nginx/user-guide/third-party-addons/opentracing/ for the most part.

I expected that like the example there, just enabling the module and providing the service name of the zipkin service (that points to my deployment) would be enough to see traces in stackdriver.

Is there another way to check whether the zipkin pod is receiving trace data? I'm trying very hard to pinpoint the issue.

[codefromthecrypt]

you can use our prometheus/grafana setup. or even ad-hoc look at the /metrics endpoint

https://github.com/openzipkin/docker-zipkin/blob/master/docker-compose.yml

On Fri, Jul 12, 2019 at 9:25 AM Yehuda Makarov notifications@github.com wrote:

That's REALLY helpful to know. thank you. No i actually don't need UI at all. But no traces are making it into stackdriver, and I'm not sure how to start trouble shooting.

I followed https://kubernetes.github.io/ingress-nginx/user-guide/third-party-addons/opentracing/ for the most part.

I expected that like the example there, just enabling the module and providing the service name of the zipkin service (that points to my deployment) would be enough to see traces in stackdriver.

Is there another way to check whether the zipkin pod is receiving any trace data?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/openzipkin/docker-zipkin-gcp/issues/8?email_source=notifications&email_token=AAAPVV5UUD6GZBS753BB4ATP67FPVA5CNFSM4IBYURAKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZYKRKQ#issuecomment-510699690, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAPVV744RUDULH4M42X7KDP67FPVANCNFSM4IBYURAA .

[yehudamakarov]

Epic! thank you! So now I'll redeploy my collector and service, and check the metrics endpoint. I guess if I'm not receiving anything from the sender, that would mean this is an issue with the ingress-nginx enable-opentracing in the configMap. Thank you for your help.

[codefromthecrypt]

no worries good luck!

On Fri, Jul 12, 2019, 9:36 AM Yehuda Makarov notifications@github.com wrote:

Epic! thank you! So now I'll redeploy my collector and service, and check the metrics endpoint. I guess if I'm not receiving anything from the sender, that would mean this is an issue with the ingress-nginx enable-opentracing in the configMap. Thank you for your help.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/openzipkin/docker-zipkin-gcp/issues/8?email_source=notifications&email_token=AAAPVV3EQ54MKOOZIBJRSR3P67GYJA5CNFSM4IBYURAKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZYK6VA#issuecomment-510701396, or mute the thread https://github.com/notifications/unsubscribe-auth/AAAPVVZYCR7N7QCUP4HWBR3P67GYJANCNFSM4IBYURAA .