HTTP Execution Queue

[kirre-bylund]

No description provided.

[kirre-bylund]

Request times have degraded slightly but within acceptable limits in my opinion:
The important numbers
Lowest 10% hovers around the same mark of 8ms±3ms (Before change was 8ms±2ms)
Average request time hovers around the same mark of 58ms±5ms (Before change was 56ms±3ms)

The degraded numbers
The highest 1% of the request has seen the largest degradation.
Before change it was 200ms±50ms, after the change it is now 300ms±50ms

Additionally, the longest running request from the test suite (which is really not a measurement of anything since it's so very circumstantial) increased from ~370ms being the highest I saw before the change to 1075ms being the highest I've seen after the change.

Conclusion
I think this is acceptable. Since the average and p10 degradation is negligible it must be the outliers that are causing the degradation. And my theory is that is almost entirely caused by me increasing max retries to 5 and applying an incremental backoff algorithm.

[kirre-bylund]

Remove AverageRequestTimeTest before merging