Reduce memory usage when queuing a lot of scans

[nabla-c0d3]

When queuing a ot lof server scans, SSLyze's memory usage increases significantly. For example, queuing 100 server scans results the SSLyze process taking 1.5 GB of memory, which seems huge. This makes it difficult to deploy SSLyze in environments with limited RAM (such as AWS Lambda).

After investigating, this is due to the fact that SSLyze relies on Python's ThreadPoolExecutor, which uses a lot of memory when submitting a large number of jobs:

• https://stackoverflow.com/questions/53104082/using-threadpoolexecutor-with-reduced-memory-footprint
• https://stackoverflow.com/questions/34770169/using-concurrent-futures-without-running-out-of-ram

[nabla-c0d3]

Fix released as part of v4.1.0.