High Availability Architecture

This document outlines the high availability design for the application infrastructure.

Goals:

• Primary hosting in Microsoft Azure
• Automatic failover to on-premise infrastructure
• Minimal downtime
• Disaster recovery backups through CBM Technology / Rader Solutions

---

1. Architecture Overview

flowchart TB

A[Architecture Overview]

A --> B[Primary Cloud Azure]
A --> C[Secondary On Prem]
A --> D[Tertiary CBM Backup]

B --> B1[Azure Application Gateway]
B --> B2[Main Application Server]
B --> B3[Primary Database]
B --> B4[External Access with IP Restrictions]

C --> C1[Replica Application Server]
C --> C2[Database Replica]
C --> C3[File Storage Replica]
C --> C4[Used During Internet Loss]

D --> D1[Daily Snapshots of Azure Primary]
D --> D2[Disaster Recovery Scenario]

---

2. Architecture Flow Diagram

flowchart LR

PC[User PC]

FW[On Prem Firewall Reverse Proxy]

VPN[Site to Site VPN]

INT[Internet]

AZGW[Azure Application Gateway]

AZAPP[Azure Application Server]

AZDB[(Primary Database Azure)]

OPDB[(Replica Database On Prem)]

OPAPP[On Prem Application Server]

PC --> FW
FW --> VPN
VPN --> INT
INT --> AZGW
AZGW --> AZAPP
AZAPP --> AZDB

AZDB --> OPDB

OPDB --> OPAPP

FW --> OPAPP

Normal operation routes traffic to Azure.

If the office internet fails, the firewall redirects users to the on-prem server.

---

3. DNS Connection

flowchart LR

PC[User PC]

GW[Gateway Firewall]

DNS[app.avexmro.com]

AZ[Azure Cloud Server]

OP[On Prem Server]

PC --> GW
GW --> DNS

DNS -->|WAN UP| AZ
DNS -->|WAN DOWN| OP

Behavior:

Normal operation

app.avexmro.com → Azure

Internet outage

app.avexmro.com → On Prem

WAN state detection is handled by the gateway firewall.

---

4. Database and File Replication

flowchart LR

AZDB[(Azure Primary Database)]

OPDB[(On Prem Replica Database)]

AZFILES[Azure File Storage]

OPFILES[On Prem File Storage]

AZDB -->|Secure Replication| OPDB

AZFILES -->|Continuous File Sync| OPFILES

Normal state

Azure → On Prem replication

Failover state

On Prem promoted to primary
Replication paused

Recovery state

On Prem → Azure resync
Azure resumes primary role

---

5. Monitoring Layer

flowchart TB

MON[Health Monitoring System]

AZAPP[Azure Application Server]
AZDB[Azure Database]

OPAPP[On Prem Application Server]
OPDB[On Prem Database]

ALERT[Alerting and Failover Trigger]

MON --> AZAPP
MON --> AZDB
MON --> OPAPP
MON --> OPDB

MON --> ALERT

Monitoring ensures:

• replication health
• server availability
• automatic failover triggers

Possible tools:

• Azure Monitor
• Zabbix
• Uptime Kuma
• Prometheus

---

6. Failover State Machine

flowchart TD

A[Normal Operation Azure Primary]

B[Internet Loss Detected]

C[Local Failover Mode On Prem Primary]

D[Internet Restored]

E[Data Resynchronization]

F[Return to Normal Azure Primary]

A --> B
B --> C
C --> D
D --> E
E --> F

Failover sequence:

1. Azure operates as primary
2. Internet outage detected
3. On-prem database promoted
4. Office operates locally
5. Internet restored
6. Data resynchronizes to Azure
7. Azure resumes primary role

---

Reliability Model

Expected behavior during failure scenarios:

Event	Result
Azure VM failure	Azure HA handles
Azure maintenance	No downtime
Office internet outage	Automatic on-prem failover
Local server failure	Cloud still available
Database corruption	Restore from CBM backup
Company power outage	Service interruption

---