How is "total documents" calculated in Sanity usage metrics?

Great question! The "total documents" number in Sanity usage is calculated with some specific rules that might explain why your GROQ query results differ.

What's Counted in Total Documents

According to the Plans and Payments documentation, the Documents quota is defined as:

"total number of stored documents (including drafts) across all datasets at the end of the month"

Important exclusions:

• System documents are NOT counted toward your quota (things like sanity.imageAsset, configuration documents, etc.)
• Assets are NOT counted as documents toward your quota
• Drafts ARE counted (they're included in the total)

Replicating the Count with GROQ

Based on the community discussion, here's how to replicate the usage count:

count(*[_type in $types])

Where $types is an array of the document types you've defined in your schema (excluding system documents and assets).

If you're getting too many documents, you're likely including:

• System documents (those with _id starting with _)
• Asset documents (sanity.imageAsset, sanity.fileAsset)

If you're getting too few, you might be:

• Missing some custom document types from your types array
• Not including drafts (draft documents have _id starting with drafts.)

Dataset Considerations

Yes, documents count across ALL datasets in your project. The quota is per-project, not per-dataset. So if you have multiple datasets (like production, staging, development), all their documents add up to your total count.

Practical GROQ Query

To get an accurate count excluding system documents and assets:

count(*[!(_id in path("_.**")) && !(_type match "sanity.*")])

This excludes:

• System documents (those in the _ path)
• Sanity's internal types (like sanity.imageAsset)

Remember that this count includes drafts, which is consistent with how Sanity calculates your usage quota.