NGINX Blue/Green Deployment not zero-downtime

[ExeErik]

We tried to build a zero-downtime deployment with your nginx canary example.

While we test the whole thing, we found out, that between the following two commands the application is not available and the nginx responds with a 404 - not found.
Between this two commands there is no nginx config within the nginx.conf for the specific hosts.

helm delete $RELEASE --namespace $DEPLOY_NAMESPACE

|
| Downtime (404)
|

helm upgrade $RELEASE $AZURE_CONTAINER_REGISTRY_NAME/multicalculatorcanary --namespace $DEPLOY_NAMESPACE --install --set replicaCount=4 --set image.frontendTag=$BUILD_BUILDNUMBER --set image.backendTag=$BUILD_BUILDNUMBER --set image.repository=$AZURE_CONTAINER_REGISTRY_URL --set dependencies.useAppInsights=true --set dependencies.appInsightsSecretValue=$APPINSIGHTS_KEY --set dependencies.useAzureRedis=true --set dependencies.redisHostValue=$REDIS_HOST --set dependencies.redisKeyValue=$REDIS_AUTH --set slot=$CANARY_SLOT --set ingress.class=nginx --set ingress.host=$INGRESS_FQDN --set canary=false --wait --timeout 45s

|
| NGINX OK (200)
|