Fixed bad pointer when function fails

[andeplane]

If function call ends up giving 503, nats-queue-worker crashes with panic: runtime error: invalid memory address or nil pointer dereference when trying to do callback.

Description

If X-Callback-Url is set, and the function call gives 503, functionRes is nil, but we try to evaluate functionRes.Header.Get("X-Duration-Seconds") in the postResult function. This leads to null pointer error and a dying pod. The call is then being retried multiple times on any new queue-worker pod starting.

This PR adds the timeout directly to the headers instead.

Motivation and Context

Crashing queue worker in the state mentioned above.

• I have raised an issue to propose this change ([required]
  queue-worker crashes if callback happens and function call failed #90

How Has This Been Tested?

I have patched the worker and tried with and without the change and verified that it does not crash with this patch. I tested this with the following images

['prom/alertmanager:v0.18.0', 'openfaas/basic-auth-plugin:0.18.11', 'openfaas/faas-idler:0.3.0', 'openfaas/gateway:0.18.13', 'openfaas/faas-netes:0.10.2-rc1', 'nats-streaming:0.17.0', 'prom/prometheus:v2.11.0', 'eu.gcr.io/cognitedata-development/queue-worker:debug2-amd64']

Types of changes

• Bug fix (non-breaking change which fixes an issue)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I've read the CONTRIBUTION guide
• I have signed-off my commits with git commit -s
• I have added tests to cover my changes.
• All new and existing tests passed.

[derek]

Thank you for your contribution. I've just checked and your commit doesn't appear to be signed-off. That's something we need before your Pull Request can be merged. Please see our contributing guide.
Tip: if you only have one commit so far then run: git commit --amend --signoff and then git push --force.

[alexellis]

LGTM