Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Timeouts - change default timeout in fetcher from 30s to 2min #454

Open
wants to merge 5 commits into
base: main
Choose a base branch
from

Conversation

timbl
Copy link
Member

@timbl timbl commented Nov 6, 2020

This is to avoid spurious timeouts called on things which in the end come though, when the net becomes slow for example, or for users on a slow link.

@timbl
Copy link
Member Author

timbl commented Nov 6, 2020

Responds to #455

@michielbdejong
Copy link
Collaborator

Hm, is this for one single http roundtrip?
When using generators this might be desirable, but when awaiting a result, 2 minutes seems like a lot.
rdflib is used in many places. We can choose any default we want, as long as it's well documented. For instance in the test-suite, I would overwrite that back to 30 seconds, probably.

@jeff-zucker
Copy link
Contributor

jeff-zucker commented Nov 18, 2020

The user can easily override the timeout setting if needed. Erring on the long side seems preferable to too short a timeout. I suppose that the timeout could be overwritten in the global solidFetcher to reset the timeout everywhere for apps that create multiple fetchers?

@michielbdejong
Copy link
Collaborator

@timbl
Copy link
Member Author

timbl commented Nov 21, 2020

"One single http round trip" isn't of course a singe IP round trip, as it can have all the TLS setup time, and CORS prefetches... it also can be delayed by the browser by being put in a pending state for a while. So when all this happens n a slow link, it is frustrating when the data actually eventually arrives but the client has given up and ignored it.

@timbl
Copy link
Member Author

timbl commented Nov 21, 2020

Yes, agreed that the fetch with the server should be set shorter ... maybe the thing to do is to cover the whole service of one incoming request, not each individual outgoing one. Would be good to to tell the client in the 500 response what it was doing when it gave up of course.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants