When trying to implement a service health check, we discovered that consul-template, a component of our service discovery mechanism, has a bug which effectively can't recognize when a service goes out of service.
The current behavior: consul-template crashes, crashing the application.
fix consul-template so that it may appropriately tell Nginx when consul recognizes that a service is available (healthy) and not.
a service should be able to cycle from healthy to unhealthy to healthy continuously. When healthy requests should be mapped to it, when healthy they shouldn't.
if two services (e.g. two deployments of the fulfillment service) are running, and one goes unhealthy, the other should be used. i.e. we should still be able to deploy as many services as we need, even if they're the same (e.g. 2 fulfillment services)
: FYI here's this ticket as we discussed. I put more detail in than likely needed. It'd be great if we can get this done if it'll only take a low LOE. Thanks!
Contract tests pass, test server working properly. Moving to Done