Closed rebeccacremona closed 6 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Comparison is base (
5986f72
) 71.11% compared to head (afe978e
) 71.12%. Report is 2 commits behind head on develop.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
When you create a batch of links via the GUI, not all your links are processed immediately: we take each job off the queue fairly, allowing other users to get captures in, in between, and your jobs are processed as our workers are able to get to them.
When we added the link batch feature, the design called for estimating how long we anticipated a given link would be queued: by taking the day's average capture time, and taking into account how many workers were available, how many links where in the queue, and any given link's position in the queue, we could estimated how long you might have to wait.
That's not worth the computational energy and DB usage: the estimation really isn't that accurate anyway.
This PR removes that feature.
If this proves distressing to users, then we have firm evidence that we should re-implement this using a different strategy in the future.
How it will look: