1. Learning the Basics of the Polly Framework

Question 1

What two categories can Polly resilience strategies be broken into?

Answer 1

1. Reactive

2. Proactive

Question 2

What is a reactive resilience strategy?

Answer 2

A resilience strategy that is triggered in response to a current and perhaps transient issue, like a server fault.

Question 3

What is a proactive resilience strategy?

Answer 3

A resilience strategy that monitors for ongoing problems and attempts to stabilise the system when facing degraded quality or longer outages in remote services.

Question 4

What reactive resilience strategies does Polly offer us?

Answer 4

1. The Retry Policy
2. The Wait and Retry Policy
3. The Circuit Breaker Policy
4. The Fallback Policy

Question 5

What does the Retry policy do?

Answer 5

Immediately retries the request some specified number of times without any delay.

Question 6

What does the Wait and Retry policy do?

Answer 6

Works like the Retry but introduces a delay before each retry.

Question 7

What does the Circuit Breaker policy do?

Answer 7

Stops all requests to an endpoint if some failure threshold is reached.

Question 8

What does the Fallback policy do?

Answer 8

Returns some specified default value if the request fails

ps. Often used in conjunction with other policies

Question 9

What can the use of delegates with the policies offered by Polly allow us to do?

Answer 9

Run custom code as part of the policy. For example, we can check if the cause of a failure was a specific error and look to resolve it before retrying the original request.

Question 10

What proactive resilience strategies does Polly offer us?

Answer 10

1. The Timeout Policy
2. The Caching Policy
3. The Bulkhead Isolation Policy

Question 11

What does the Timeout policy do?

Answer 11

Set the timeout for a request rather than waiting for default as implemented by the HttpClient.

Question 12

What does the Caching policy do?

Answer 12

Cache responses for a period of time and then channel future requests to the cache rather than the remote service.

Question 13

What does the Bulkhead Isolation policy do?

Answer 13

Manage the load on your own application by limiting the number of concurrently executing requests and limiting the size of the queue of requests that are waiting to execute.

Question 14

What is a resilience framework?

Answer 14

There is no agreed-upon definition, but you can think of it as a set of libraries that help an application to recover from transient or longer failures in services or infrastructure upon which it depends when recovery is possible, and that facilitates graceful degradation of your application when it is not.

Question 15

When does a Retry policy retries the request?

Answer 15

Immediately after failure.

Question 16

What does the Wait and Retry policy allow us to define?

Answer 16

Backoff between entries, allowing us to increase the backoff of each subsequent request.

Question 17

How does the Wait and Retry policy work for handling a request that failed?

Answer 17

It will decide whether to retry.

1. If it decides to retry, it waits a specified amount of time before tries to send the request again.
2. If no retries should be performed, it returns the response.

Question 18

What parameters do the WaitAndRetry method take?

Answer 18

It takes the number of times to retry; and a lambda function that takes in an int parameter which tells us the retry count and returns a TimeSpan that determines how long to wait before retrying.

...
    .WaitAndRetryAsync(3,
        retryAttempt => TimeSpan.FromSeconds(
             Math.Pow(2, retryAttempt) / 2));

Question 19

What benefit does the Wait and Retry policy offer?

Answer 19

It gives the remote service time to recover, and in some scenarios, will make the retry more likely to succeed.

Question 20

What is idempotence in the context of web requests and how does it relate to Polly policies?

Answer 20

A request is idempotent if it can be called any number of times and the result is the same each time.

GET, PUT, and DELETE are usually designed to be idempotent, but you should not expect POSTs to be. So, be careful when retrying POSTs.

Question 21

What does the use of delegates allow us to do when used with policies?

Answer 21

It allows us to execute arbitrary code before the request is retried — log a failure, call some other method, etc.

/...
.RetryAsync(3,
   onRetry: (httpResponseMessage, _) =>
   {
      if (httpResponseMessage.Result.StatusCode == HttpStatusCode.Unauthorized)
      {
         PerformReauthoripzation();
      }
   });

Question 22

What does the Fallback policy do?

Answer 22

It allows us to return some meaningful default if all else has failed.

Question 23

How does the Fallback policy work when used in conjunction with a HTTP request?

Answer 23

If the request succeeds, the response is returned as normal. If the request fails, the Fallback returns some default response.

Question 24

What’s a common policy to use in conjunction with the Fallback policy?

Answer 24

The Retry policy — if the retries fail, use the Fallback to return a default.

Question 25

What does the Fallback policy offer us when it comes to executing arbitrary code?

Answer 25

Like Retry and Wait and Retry, Fallback offers a delegate that can be called before the Fallback’s behavior clause is executed.

Question 26

When using the HttpClient to call a remote service, what will it do if the remote service doesn’t respond within a certain time?

Answer 26

Throws an HttpRequestException.

Question 27

How does the workflow for simply handling a HttpClient timeout with Polly look like?

Answer 27

A request is made. If a HttpClient timeout occurs, we check if we should perform a retry. If we should, we make the request again; if we shouldn’t, we’ll propagate that timeout exception.

Question 28

What is the default time to timeout when making a request with the HttpClient?

Answer 28

100 seconds.

Question 29

How does the Timeout policy work?

Answer 29

1. You specify a period you’re willing to wait for a request to respond.
2. You send that request
   3a. If no response is received within the specified time, the Timeout policy throws a TimoutRejectedException. At this point, you can also pass a cancellation token to the HttpClient, allowing it to cancel the request and release any resources it’s holding.
   3b. If you did receive a response within the specified time period, normal operation resumes and the response is returned to the caller.

Question 30

What methods of the Policy type can we use to set a timeout policy in Polly and what parameters do they take?

Answer 30

The Timeout or the TimeoutAsync method. They take an integer parameter indicating the number of seconds to wait before timing out.

var policy = Policy.TimeoutAsync(1);

Question 31

What methods of the Policy type can we use to set a fallback policy in Polly and what parameters do they take?

Answer 31

The Fallback or the FallbackAsync method in conjunction with the HandleResult method. It takes a type T parameter for the default to return.

var p = 
Policy.HandleResult(
        r => !r.IsSuccessStatusCode)
    .FallbackAsync(
        // some default to return...
    )

Question 32

What methods of the Policy type can we use to set a retry policy in Polly and what parameters do they take?

Answer 32

The Retry or the RetryAsync method in conjunction with the HandleResult method. It takes in an integer parameter indicating the number of times to retry the request before giving up.

Policy.HandleResult(
        r => !r.IsSuccessStatusCode)
    .RetryAsync(3);

Question 33

How do you run a Polly policy and what argument must you supply in order to do so?

Answer 33

By calling the Execute method for synchronous policies or the ExecuteAsync for the asynchronous ones. You supply it with an anonymous function to execute, which must return the type specified in the HandleResult clause.

policy.ExecuteAsnyc(
    () => httpClient.GetAsync(requestEndpoint));