Initial draft

This project is an effort to mimic a streaming analytics pipeline. In general, this pipeline will stream data from OLTP database to OLAP database, which has the following details:

1. OLTP: Postgresql will serve as a producer. A replication mechanism can be implemented to ensure the high availability - if the primary server fails, one of the stand-bys can take over. There are two kinds of replication in Postgresql: Streaming Replication and Logical Replication.
2. Change Data Capture (CDC): To capture any change on the upstream source, we will use CDC technique. Debezium with Kafka - Postgresql connector will be applied here. Debezium is widely regard as an industrial standard for CDC mechanism.
3. Streaming Processor and/or Database: Flink (streaming processor) or SparkStructuredStreaming (streaming processor), RisingWave (streaming database) are our top options. (This part will be considered after I finish the OLTP and CDC parts. Note that there are also many other options besides these three).
4. OLAP: Clickhouse is our OLAP database here. Clickhouse is one of the most popular OLAP database in the market, it also has OSS version with detailed documentation, which makes our implementation easier.
5. Observability: This pipeline should have one but first, let me finish 4 previous parts.