Availability and scaling. Http paths for checking readiness and liveness.

We offer to use next URLs for checking health of the service

/maintenance/health_check will return 2xx Http status if web server is reachable. You may put this path into docker-based cluster orchestrator like Kubernetes to check service can response to requests to manage traffic across instances (ex: Kubernetes Pods)
/maintenance/liveness_check will return 2xx Http status if application is configured and all dependency services are reachable and successfully configured. You may put this path into orchestrator like Kubernetes to check running container is live. This handler will check dependency services are running successfully.

Scaling and dependencies:

Postgres: persistence component, application don't route read sql queries to replicas. To avoid this restriction use tools like Pgpool-II
Memcached: caching component. To make memcached high-available you may deploy it to Cluster as service or use alternate approaches like this
Redis: Use HA Redis cluster or deploy multiple redis instances as separate services (with scale replicas = 1) in a cluster (cluster orchestrator will schedule instances for HA purposes).