- Add new Ansible role `wireguard` to set up WireGuard VPN server on
DB node with key generation, firewalld rules, and client peer config.
- Introduce `pg-proxy` and `mongo-proxy` socat containers in db_stack
to expose PostgreSQL (15432) and MongoDB (17017) on host ports,
restricted to WireGuard subnet via firewalld.
- Update test environment group_vars with WireGuard client entry for
`murat-inspiron-15-3525`.
- Modify act_runner config: set `docker_host` to unix socket, remove
explicit socket mount from options, and change runner label image to
`catthehacker/ubuntu:act-22.04`.
- Open UDP port 51820 in Hetzner firewall for WireGuard inbound.
- Adjust test-db-post-stack playbook to include wireguard role (tagged).
- Update roadmap document with APISIX init step order.
This commit systematically updates all Terraform configurations, including resources, variables, and labels, to use the more generic `app` designation instead of `swarm`. This improves consistency and decouples the infrastructure naming from a specific container orchestration technology like Docker Swarm.
Overhaul and expand firewall definitions for both `prod` and `test` environments to enable comprehensive inter-subnet communication.
This includes implementing explicit rules supporting:
- Docker Swarm overlay networks between application and database subnets.
- High-availability database clusters (PostgreSQL replication, MongoDB replica sets, Patroni, etcd).
- Internal access for various infrastructure services (Vault, Redis, RabbitMQ, APISIX, Prometheus, Grafana).
All firewall rule descriptions are standardized in English for improved clarity and consistency.
Additionally, update default `server_type_swarm` and `server_type_db` variables to the recommended `CPX` series for both environments. An initial generated Ansible inventory for the test environment is also added.
- This commit introduces the Terraform configuration to provision a production environment on Hetzner Cloud, building on the existing test setup.
- Key improvements and new features include:
* **Multi-node clusters:** Scaling to 3-node Swarm application and database clusters for improved resilience.
* **High availability:** Utilizing a Hetzner Floating IP for the application entry point and `spread` placement groups for fault tolerance across physical hosts.
* **Enhanced network security:** Internal management services (RabbitMQ, APISIX, Prometheus, Grafana) are restricted to the application subnet, expected to be accessed via an internal reverse proxy (SWAG).
* **Internal database replication:** New firewall rules enable PostgreSQL replication and MongoDB replica set traffic within the database subnet.
* **Refined test environment:** Updates to align `test` configuration with the new `prod` structure, including a dedicated floating IP and adjusted firewall rules.
* **Configuration standardization:** Environment-specific details moved to `locals.tf` for clarity, with upgraded server types and migration to Rocky Linux as the base image.
- Updates were also made to the latest version of Terraform to ensure consistency in the documentation
This commit introduces the foundational Infrastructure-as-Code for provisioning a test environment on Hetzner Cloud. It defines server nodes, private networking, comprehensive firewalls, and includes documentation on resource lifecycle and safe configuration practices.
