Diagnose and speed up slow or expensive Apache Spark jobs using the Spark History Server MCP tools — finds the real bottleneck (skew, spill, shuffle, small files, GC, bad join strategy, under-parallelism) from task-level metrics and returns ranked, quantified fixes. Use this whenever someone asks why a Spark job is…
Root-cause a failed, killed, or hung Apache Spark application using the Spark History Server MCP tools — reads the actual task exceptions, stage metrics and runtime config, then names the cause and the fix. Use this whenever someone mentions a Spark job that failed, crashed, errored, was killed, OOMed, got stuck, is…