depesz

explain.depesz.com now shows two IO bandwidth metrics for parallel query nodes: average per-worker bandwidth and total exclusive-time disk bandwidth. Previously, only the per-worker average (~273 MB/s) was displayed. The new view also shows the aggregate disk throughput required during wall-clock execution time (~739 MB/s in the example), making it easier to understand actual disk pressure during parallel scans where timing data is summed across all workers.

Per-worker, and global, IO bandwidth in explain plans – select * from depesz;