High availability design patterns in Azure - Step-by-Step Execution

Process Flow - High availability design patterns

Start: User Request

↓

Load Balancer receives request

↓

Check service instance health

↓

Route to

↓

Healthy

↓

Instance

↓

Retry or Failover

↓

Process Request

↓

Send Response

↓

End

User requests go through a load balancer that checks service health and routes to healthy instances, retrying or failing over if needed to keep services available.

Execution Sample

Azure

1. User sends request
2. Load balancer checks instance health
3. Routes to healthy instance
4. Instance processes request
5. Response sent back

This flow shows how a request is handled to ensure high availability by routing only to healthy service instances.

Process Table

Step	Action	Instance Health Check	Routing Decision	Result
1	User sends request	N/A	N/A	Request received by load balancer
2	Load balancer checks instance A	Healthy	Route to instance A	Request sent to instance A
3	Instance A processes request	N/A	N/A	Request processed successfully
4	Response sent back to user	N/A	N/A	User receives response
5	Next request arrives	Instance A unhealthy	Route to instance B	Request sent to instance B
6	Instance B processes request	N/A	N/A	Request processed successfully
7	Response sent back to user	N/A	N/A	User receives response
8	Instance B unhealthy	Failover to instance C	Route to instance C	Request sent to instance C
9	Instance C processes request	N/A	N/A	Request processed successfully
10	Response sent back to user	N/A	N/A	User receives response
11	No healthy instances	All unhealthy	Fail request or queue	Request fails or delayed
12	End	N/A	N/A	High availability maintained by routing or failover

💡 Execution stops when request is successfully processed or no healthy instances remain.

Status Tracker

Variable	Start	After Step 2	After Step 5	After Step 8	Final
Instance A Health	Healthy	Healthy	Unhealthy	Unhealthy	Unhealthy
Instance B Health	Healthy	Healthy	Healthy	Unhealthy	Unhealthy
Instance C Health	Healthy	Healthy	Healthy	Healthy	Healthy
Current Routed Instance	None	Instance A	Instance B	Instance C	None or last healthy
Request Status	Pending	Routed	Routed	Routed	Completed or Failed

Key Moments - 3 Insights

Why does the load balancer route to a different instance after step 5?

Because instance A became unhealthy at step 5, the load balancer detects this and routes the request to instance B to maintain availability, as shown in execution_table row 5.

What happens if all instances are unhealthy?

When all instances are unhealthy, as in step 11, the system cannot route requests successfully, so requests fail or are delayed, ensuring no routing to unhealthy instances, as shown in execution_table row 11.

How does the system ensure the user still gets a response if one instance fails?

The load balancer checks health and routes to another healthy instance automatically, retrying the request, as seen in steps 5 to 10 in the execution_table.

Visual Quiz - 3 Questions

Look at the execution_table, at which step does the load balancer first route to instance B?

A Step 2

B Step 5

C Step 8

D Step 11

💡 Hint

Check the Routing Decision column in execution_table rows around step 5.

According to variable_tracker, what is the health status of instance C after step 8?

A Healthy

B Unhealthy

C Unknown

D Pending

💡 Hint

Look at the 'Instance C Health' row and the 'After Step 8' column in variable_tracker.

If instance A remained healthy, how would the routing change in the execution_table?

A Requests would fail at step 11

B Requests would route to instance B first

C Requests would always route to instance A

D Load balancer would route randomly

💡 Hint

Refer to the routing decisions in execution_table when instance A is healthy.

Concept Snapshot

High availability means keeping services running without interruption.
Use load balancers to check instance health.
Route requests only to healthy instances.
If one fails, failover to another.
If none healthy, requests fail or queue.
This pattern keeps apps available to users.

Full Transcript

High availability design patterns in Azure use load balancers to manage user requests. When a user sends a request, the load balancer checks which service instances are healthy. It routes the request only to healthy instances. If an instance becomes unhealthy, the load balancer redirects requests to other healthy instances to keep the service running. If all instances are unhealthy, requests may fail or be delayed. This ensures users get responses reliably by avoiding failed instances and using failover automatically.