My advice would be deep dive on the technology / libraries you use. For instance I know MySQL fairly well. It's known to be one of the less optimized SQL databases but because I know many pitfalls and optimizations I can use MySQL much more efficiently than any other SQL db.
Tons of people switch from technology X to technology Y expected a huge performance gain because some blog post said they got a huge gain. However, if you don't know the new technology you are bound to use it incorrectly and can often end up with worse performance.
This also applies to libraries. My favorite examples are Java HTTP Clients. Most people copy paste the working code from the readme and move on. What they don't realize is that most Java HTTP clients limit you to a very small number of concurrent HTTP connections to the same hostname / route. For example in Apache HTTP Client its 2 concurrent connections and OkHttp I think its 5. In microservice land where you generally only hit a few hosts you now severely limited your throughput depending on the service. Simply tweaking a config variable can give you way more performance gains than switching libraries if you just do your research.