ValidationException: Item size has exceeded the maximum allowed size

TL;DR — A DynamoDB item can be at most 400 KB (attribute names + values combined). Your write pushes an item over that. Move the large field out (to S3, or split across items) and store a reference instead.

What it means

ValidationException: Item size has exceeded the maximum allowed size of 409600 bytes

409,600 bytes = 400 KB. The limit counts the entire item: every attribute name plus its value, UTF-8 encoded, including nested map/list overhead. An UpdateItem that grows an existing item past 400 KB fails the same way.

Why it happens

Storing large blobs inline — base64 images, PDFs, big JSON documents.
An unbounded list/map (append-only arrays, event logs) that grows over time until it crosses 400 KB.
Long attribute names multiplied across a big item.
Denormalizing too much into a single item.

How to fix it

Offload large values to S3. Store the object in S3 and keep only the key/URL in DynamoDB. This is the standard pattern for anything approaching the limit.
Split the data across multiple items. Use the item-collection / vertical-partition pattern — one logical entity as several items sharing a partition key.
Cap growing collections. Don't let a single item accumulate an unbounded list; roll entries into child items keyed by sort key.
Compress genuinely large text before storing (gzip → binary attribute), if S3 isn't an option.

Example — reference pattern

// Instead of storing the blob inline, store an S3 pointer:
await doc.send(
  new PutCommand({
    TableName: 'Documents',
    Item: {
      pk: 'DOC#1',
      title: 'Q3 report',
      s3Key: 'documents/DOC#1/report.pdf', // the bytes live in S3
      sizeBytes: 2_400_000
    }
  })
);

ItemCollectionSizeLimitExceededException — the 10 GB collection limit (LSI tables).
ValidationException (overview)
Learn: Item size & the 400 KB limit · Item collections