Enabling compression for fields, properties, and indexes conserves storage space but can significantly affect query performance. It is advisable to assess the impact of data compression before implementation.
Relevant recommendations:
Avoid enabling compression for frequently accessed indexes, fields, and properties;
Refrain from compressing hot spot fields;
Compression can significantly impact query performance. Conduct grayscale testing in a test cluster to gauge the effects of data compression on performance metrics such as query latency and cluster load. If the business impact is acceptable, consider updating the online serving cluster.
Optimization methods after enabling data compression:
Despite the notable performance impact, enabling compression can be mitigated by optimizing the index table's loading policy. The configuration is as follows:
In the cache policy, add the following parameter to load_config>load_strategy_param:
"cache_decompress_file": trueThis setting caches compressed data, which, when used in tandem with data compression, can help mitigate performance degradation.