Backoff to avoid excessive retries to Run Service in a duration #3354
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ericsciple
commented
Jun 21, 2024
| @@ -62,7 +62,10 @@ public Task<AgentJobRequestMessage> GetJobMessageAsync(string id, CancellationTo | |||
| CheckConnection(); | |||
| return RetryRequest<AgentJobRequestMessage>( | |||
| async () => await _runServiceHttpClient.GetJobMessageAsync(requestUri, id, VarUtil.OS, cancellationToken), cancellationToken, | |||
| shouldRetry: ex => ex is not TaskOrchestrationJobAlreadyAcquiredException); | |||
| shouldRetry: ex => | |||
There was a problem hiding this comment.
The problem with not retrying these errors, means we immediately call to Broker to get the next message, which might be the same job message.
When one of these errors is encountered from Run Service, an async message is sent to Broker. We don't want to keep hammering the service in a tight loop if there are delays processing the message queue.
| maxBackoff: s_maxBackoff, | ||
| deltaBackoff: s_backoffCoefficient); | ||
| Trace.Warning($"Back off {backoff.TotalSeconds} seconds before next attempt."); | ||
| await Task.Delay(backoff, token); |
There was a problem hiding this comment.
FYI, Task.Delay will throw on cancellation token fire.
There was a problem hiding this comment.
I think that's OK, but I switched to Host.Delay so I could unit test this
ericsciple
force-pushed
the
users/ericsciple/24-06-backoff
branch
from
fb896c0
to
d73aabc
Compare
ericsciple
force-pushed
the
users/ericsciple/24-06-backoff
branch
from
d73aabc
to
ec1e338
Compare
TingluoHuang
approved these changes
Jun 24, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Helps avoid excessive calls to Run Service when encountering non-retriable errors from /acquirejob. Normally we rely on the HTTP clients to back off between retry attempts. However, acquiring a job involves calls to both Run Serivce and Broker. And Run Service and Broker communicate with each other in an async fashion.
When Run Service encounters a non-retriable error, it sends an async message to Broker. The runner will, however, immediately call Broker to get the next message. If the async event from Run Service to Broker has not yet been processed, the next message from Broker may be the same job message.
The error throttler helps us back off when encountering successive, non-retriable errors from /acquirejob.