Hello,
We recently had a surge in following issue, although this seems like something that always existed (With Apollo iOS 0.x too).
We implement urlSession(_:task:didFinishCollecting:) delegate method and collect request metrics, We saw URLSessionTaskTransactionMetrics object (transactionMetrics.first) sometimes reports a negative request duration. Here we calculate request timing as the interval between metrics.responseEndDate and metrics.fetchStartDate and for some requests, this value is negative.
This issue occurs randomly, across different query types and endpoints and Our URLRequest.cachePolicy for most of these requests is set to .reloadIgnoringLocalCacheData for most of these graphql requests.
Are we aware of this issue or what might cause request timings to be negative? This might be fixed in Apollo 2.0 with networking stack changes but I want to understand what is happening here.
Thanks!