Describe your cluster and job and get tuned Apache Spark settings — executor cores and memory, executor count, driver memory, shuffle partitions and overhead — with the reasoning behind them.
Tell us your node count, cores per node and memory per node.
Explain the workload — data sizes, joins, shuffle intensity and output.
Get concrete spark-submit / SparkConf values plus the reasoning, then apply and benchmark.
Yes, it's free and needs no signup. A fair-use rate limit keeps it available for everyone.
Executor cores and memory, number of executors, driver memory, memory overhead, shuffle partitions, and relevant tuning flags — with justification.
No. It gives strong, well-reasoned starting values. Always benchmark on your real data and adjust, since optimal settings depend on your workload.
