Problem: The BigchainDB Server to Tendermint-broadcast-API connection gets flaky in the long run

[ttmc]

This issue was noticed by @vrde when doing long-running performance tests.

Potential Solutions:

• Do something to make it more reliable.
• Give users the option to post new transactions directly to the Tendermint broadcast API, via WebSocket. That should be more reliable. Document how they can do that. Expalin the reliability, security, and UX tradeoffs. Add support to drivers?

[shahbazn]

@vrde is there any update on this issue? I would like to understand the exact issue and possibly push a fix

[ttmc]

This issue might be related to issue #2292