[enhancement] OzHera supports multiple data sources

[sadadw1]

Due to the constraints of open-source licenses and in consideration of flexibility and scalability, we are contemplating adapting to and supporting multiple data sources.

1、We are considering switching the scenarios of reading and writing data in OzHera from Elasticsearch (ES) to Doris or ClickHouse. This requires us to make some changes:

Ideally, we should encapsulate the reading and writing of data sources into a common SDK. This way, the SDK can be used directly in the relevant places. The benefits of this approach include:

1. no need to introduce unnecessary dependencies in the calling project.
2. minimal code modifications in the calling project.

All instances where ES is used need to be modified to call the encapsulated SDK. This includes:

1. ozhera-log
2. trace-etl
3. ozhera-mimonitor

2、We are also considering supporting various relational databases, not just MySQL, such as PostgreSQL. This may require us to modify all projects that use relational databases, including changes to POM files and configurations.