Front Door Load balancing

Nibbler 616 Reputation points
2024-06-01T08:51:18.0566667+00:00

Hello everyone,

I am looking for detailed information on how Azure Front Door load balances between origins. I have found some helpful information at https://learn.microsoft.com/en-us/azure/frontdoor/routing-methods, but I still have some questions.

Assuming there are two origins available (A+B), and both have a priority of 1, latency sensitivity of 0, and a weight of 1000:

  1. How will traffic be distributed?
  2. If origin A becomes unhealthy, will B take over?

I would appreciate any insights the community can offer. Thank you.

Maybe you @GitaraniSharma-MSFT take on this?

Azure Front Door
Azure Front Door
An Azure service that provides a cloud content delivery network with threat protection.
615 questions
0 comments No comments
{count} votes

2 answers

Sort by: Most helpful
  1. Q&A Assist
    2024-06-01T08:51:18.0966667+00:00
    AI generated content. This question contains an answer created with AI from Q&A Assist. Learn more

    Azure Front Door supports four different traffic routing methods to determine how your HTTP/HTTPS traffic is distributed between different origins: Latency, Priority, Weighted, and Session Affinity. Assuming there are two origins available (A+B), and both have a priority of 1, latency sensitivity of 0, and a weight of 1000, traffic will be distributed evenly between the two origins. If origin A becomes unhealthy, Azure Front Door will automatically route traffic to origin B.


    References:

    0 comments No comments

  2. KapilAnanth-MSFT 38,051 Reputation points Microsoft Employee
    2024-06-03T06:42:41.1333333+00:00

    @Nibbler ,

    Welcome to the Microsoft Q&A Platform. Thank you for reaching out & I hope you are doing well.

    I understand that you would like to understand how AFD Traffic routing methods work.

    We should follow the Overall decision flow chart

    • User's image

    1.How will traffic be distributed?

    1. Available origins : A and B are available, so both A and B are selected
    2. Priority : A and B have same priority of 1, so both A and B are selected
    3. Latency signal (based on health probe) :
      1. Let us assume Latency sensitivity of 0 (0 is highly unlikely, but we shall go with it for this example) and origin group's latency sensitivity is set to 30 ms.
      2. Since 0ms is less than 30ms, both A and B are selected.
    4. Weights: A and B both have a weight of 1000.
      1. This means, the total is 2000 (A+B)
      2. And then traffic distributed to A is 1000/2000 (A/Total) , which is 1/2
      3. And traffic distributed to B is 1000/2000 (B/Total) , which is 1/2

    i.e., Traffic is distributed equally across A and B

    2.If A becomes unhealthy,

    1. Available origins : Since A is unhealthy, only B is selected
    2. Since there is only one Origin now, no need for further validation.

    i.e., 100% Traffic is routed to B

    Kindly let us know if this helps or you need further assistance on this issue.

    Thanks,

    Kapil


    Please don’t forget to close the thread by clicking "Accept the answer" wherever the information provided helps you, as this can be beneficial to other community members.

    0 comments No comments