Troubleshooting
Docker
Restart policy in a swarm
When the docker-swarm-production.yml swarm services have on-failure configured and the service fails to start up and become healthy, it'll get stuck in an infinite restart loop and block all other services from starting up. In a swarm, it's best to let the swarm handle all services and only set a delay.
See https://stackoverflow.com/a/56564483/2362944
Stack rollbacks
When you publish changes to a stack and there's at least one failure, the swarm will roll back those changes and revert to the old configuration.
However, the deployment recipe requires only a few essential services to be healthy before continuing with the release and this can either be with the new or rolled back configuration.
Redis .env configuration
A DigitalOcean managed Redis database only allows encrypted connections, so you need to explicitly enable that in the host variable by prefix the URL with tls://:
REDIS_HOST=tls://smarthealth-app-trial…References: