Check for git hash on deployments

[teneightfive]

Problem

Our current deployment job within Circle waits for the command to trigger the kubernetes deployment to finish before marking it as complete and triggering the rest of the workflow.

This has caused false positives to occur when a deployment was not successfully completed and the pods did not start with the correct code deployed.

Possible solution

To ensure a correct positive we could poll the healthcheck endpoint with the git sha that is expected to be deployed to wait for the pods to have successfully started before completing the job.