default MDS shard size; 50–100 MB shards work well across modalities (large sequential reads, low metadata load)

Name: default MDS shard size; 50–100 MB shards work well across modalities (large sequential reads, low metadata load)
License: https://aidatacenterguide.com/terms#intellectual-property

~67 MBobserved

Value kind	observed — Reported measurements, counts, and specifications keep the precision and scope stated by their source; an exact specification is not treated as a range.
As of	2025
Source	Databricks Mosaic Streaming docs · “Don’t make your shard file size too large or small”: default 67 MB; 50–100 MB works well
Review	checking…review by 2026-09-05 · standard cadence
Recorded changes	last 2026-06-29
Claim id	default-mds-shard-size-50-100-mb-shards-work

Where the guide uses it

9.5 Data Ingestion, Preprocessing & the Data-Loader Path

← Full numbers register — every date-stamped figure in the guide, with revision history.