Reduce the default maxChainBytesPerSyncAttempt to 50MB

[Wayonb]

Describe the bug
During the sync of testnet, the catapult-client will timeout when pulling blocks. This is because certain block are have a lot of transactions and on slower networks takes a while to pull 100MB.

Should lower the maxChainBytesPerSyncAttempt to 50MB

To Reproduce
To reproduce the behavior:

1. Try to sync testnet

[dusanjp]

With the settings created by the current bootstrap, it is difficult to exceed 130681 blocks with testnet (except for nodes with abundant cpu resources).
After this 130681 block, it is the time when nft-drive started test operation.
Furthermore, we are conducting a stress test at the end of August, but it is also a heavy block here, which is hindering synchronization. It is about 330000 blocks or later.

Here, I tried with the following settings based on the advice of those who noticed it:
maxChainBytesPerSyncAttempt: 10MB
messageSynchronizationMaxResponseSize: 5MB
blockDisruptorMaxMemorySize: 1000MB

If node is stopped for a long time and then restarted or a new node is started without making the above settings, synchronization could not be completed (except for nodes with abundant cpu resources).

The default values for these settings are:
maxChainBytesPerSyncAttempt: 100MB
messageSynchronizationMaxResponseSize: 20MB
blockDisruptorMaxMemorySize: 300MB

Since nft-drive will be running on mainnet in the near future
In the next bootstrap, I think it is necessary to adjust the default values of these values.