Earth observation technology has become increasingly crucial for monitoring various aspects of our planet through systematic data collection. However, processing the large volume of satellite data generated can be challenging, requiring high-performance computing solutions. The Dask framework has gained popularity for its flexibility and efficiency in processing large amounts of data in a distributed manner. Nevertheless, determining the optimal Dask cluster configuration remains challenging, as it requires balancing performance and cost objectives. A novel multi-objective optimization service is proposed to address this challenge that enhances the performance and cost efficiency of earth observation data processing workflows. Our approach is to generate a set of Pareto-optimal solutions, allowing users to make informed decisions regarding the optimal trade-offs between performance and cost. The effectiveness is demonstrated using real-world earth observation datasets and outperforms existing performance and cost-efficiency solutions.