fix(aws-lambda): reduced latency for large number of spans

[kirrg001]

refs https://jsw.ibm.com/browse/INSTA-13498

• live test with lambda layer
• live test without lambda layer
• can the backend connect the spans still? If we send 10 each?
  • backend confirmed: no big deal as long as they arrive within 20s.
  • add a protection for the 20s limit!
• layer too slow with AND without span buffering - https://github.ibm.com/instana/lambda-extension/pull/18
• create multiple PR's

Span buffering = Just send a batch of requests as soon as you see more than X.

100 http requests with 100ms delay

The results are WITHOUT cold start. Cold start is always a bit shitty for everything.

Without layer, without Instana

• 10066
• 10152
• 10031
• 10060
• 10231
  - avg: 10,108

Without layer, with Instana, without span buffering

• 10317
• 10343
• 10171
• 10401
• 10192
• you clearly see long and short latencies
  - avg 10,284

Without layer, with Instana, with span buffering

• 10312
• 10203
• 10215
• 10192
• 10187
• 10194
  - avg 10,183

with layer, without instana

• 10208.08
• 10030.98
• 10027.15
• 10031.04
• 10038.49
  - avg 10047.35

with OLD layer, with Instana, without span buffering (256mb)

• 10555.34 ms
• 10664.37 ms
• 10354.12 ms
• 10372.81 ms
• I stopped doing more tests, because the results were too bad
  - avg 10486.66

with OLD layer, with Instana, with span buffering (256mb)

• 10633.13 ms
• 10489.61
• 10287.26 ms
• 10866.70 ms
• 10481.99
• I stopped doing more tests, because the results were too bad
  - avg 10551.74

with new layer, with Instana, without span buffering (256mb)

• 10343.37
• 10214.14
• 10172.68
• 10181.04
• 10163.27
• 10163.64
• 10170.47
• 10171.97
• 10172.07
• 10169.18
  - avg 10182.24

with new layer, with Instana, with span buffering (256mb)

• 10120.98
• 10139.68
• 10426.91
• 10176.06
• 10185.97
• 10116.41
• 10173.31
• 10125.05
• 10177.48
• 10210.57
• 10130.88
  - avg 10,139.42