Cloudflare GraphQL queries must select fields the dataset exposes

Reject CF GraphQL queries that select fields the dataset doesn’t expose. The two failure modes are symmetric:

Catches the same class of footgun as app-correctness/cf-graphql-dataset-by-window but for the schema axis rather than the time window axis. The two compose: one rules out wrong-dataset-for-window, the other rules out right-dataset-wrong-fields.

Schema cheat sheet#

Note that dimensions, avg are valid in both - the asymmetry is count vs sum.

Fix#

# WRONG - count on 1dGroups doesn't exist
httpRequests1dGroups(filter: { date_geq: "2026-05-11" date_leq: "2026-05-18" }) {
  count
}

# RIGHT
httpRequests1dGroups(filter: { date_geq: "2026-05-11" date_leq: "2026-05-18" }) {
  sum { requests }
}

# WRONG - sum on Adaptive doesn't exist
httpRequestsAdaptiveGroups(filter: { datetime_geq: "..." datetime_leq: "..." }) {
  sum { requests }
}

# RIGHT
httpRequestsAdaptiveGroups(filter: { datetime_geq: "..." datetime_leq: "..." }) {
  count
}

Suppressing intentional cases#

For one-off historical queries explicitly using a non-standard field combination (e.g. on a paid tier that exposes additional fields):

# appframes:disable-next-line app-correctness/cf-graphql-schema-match
httpRequestsAdaptiveGroups(...) {
  sum { requests }   # custom paid-tier extension
}

Generalizes to#

Any analytics API with multiple datasets with overlapping but non-identical schemas:

• GA4 dimensions vs metrics per report type
• Datadog query types (metrics vs events vs logs)
• BigQuery views over underlying tables (column subset can change)
• Snowflake share schemas

The frame currently codifies Cloudflare’s two main dataset families. Adding others requires populating cfDatasetSchemas with the equivalent valid/invalid field maps.