Switch to event and store listening

[RiccardoM]

Context

Currently, in order to listen for state changes, BDJuno uses the following methods:

1. inside BDJuno v1, the changes are fetched using a pull strategy: the RPC/gRPC endpoints are queried periodically (each block or based on a timer)
2. inside BDJuno v2, the changes are fetched either as in v1, or by querying the local storage.

While these methods have worked until now, they have become increasingly harder to maintain. We have particularly seen some of their limitations inside chains that have a huge amount of users like the Terra Chain. These chains make it extremely harder to periodically fetch some useful data such as the delegator rewards amount. Due to the high amount of data to get, the gRPC can become very slow to respond making the entire node loose blocks and become out-of-sync with the chain itself.

For this reason, I think we should try to leverage the new features that are being developed from both the Tendermint and Cosmos teams:

• Tendermint PostgreSQL indexer for events which stores each transaction event inside a PostgreSQL database using the following tables.
• Cosmos ADR 038 - State listening, which will allow Cosmos chains to stream the state changes to files or to an auxiliary streaming service over gRPC which can be accessed by BDJuno either locally or over the network. A track of this ADR implementation status is kept here: Postgres Indexing (ADR038) Implementation cosmos/cosmos-sdk#9817