Add identity gdpr erasure

[jamessnowplow]

What changed?

Added identity erasure docs.

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	documentation	2c65693	Commit Preview URL Branch Preview URL	Aug 21 2026, 09:45 AM

[claude]

Documentation style review

A few items to address — mostly minor, plus one merge-blocker the author already flagged.

1. "fulfil" → "fulfill" (US spelling)

"to fulfil a GDPR right-to-erasure request"
"to fulfil GDPR right-to-erasure requests"
"To fulfil the request, you must also delete..."
description: "Erase a user's entries from the Snowplow Identities graph to fulfil a GDPR right-to-erasure request."

CLAUDE.md → Writing style: "Use US English spelling." The US form is fulfill (double L). Replace each instance in docs/identities/erasure/index.md (intro, warning admonition, "Make an erasure request" intro, and the frontmatter description).

2. Bold used for non-UI emphasis

• 100 combined inputs (identifiers plus snowplow_ids)
• 256 KB request body

CLAUDE.md → Grammar and formatting: "Use bold only for UI elements (buttons, page titles)." The style guide reinforces this (bold = buttons, page titles, tabs, menus). These are values, not UI elements — drop the bold:

- 100 combined inputs (`identifiers` plus `snowplow_ids`)
- 256 KB request body

3. Pending "do not merge" gate

{/* TODO: do not merge this page to main until Console > Identities > Overview surfaces the API URL referenced in the Authenticate section. */}

Outside pure style scope, but worth surfacing: the author has marked the page as not-ready-to-merge until Console exposes the API URL the Authenticate section sends readers to find. Either resolve that dependency or convert this into a draft/hold before merge. Also remove the comment once the gate clears.

4. Minor: heading voice

## Fulfilling GDPR erasure requests (in docs/identities/index.md)

CLAUDE.md → Headings and organization: "Prefer imperative voice." Consider ## Fulfill GDPR erasure requests (and fix the spelling per item 1).

---

Otherwise the page is well-structured: frontmatter is complete, internal links use absolute paths ending in /index.md, H2 count is within range, terminology (entity, identity graph, lowercased concepts) matches the guide, and the request/response examples are consistent between dry-run and real-erasure sections. No external URLs were added that needed link-checking.

[greptile-apps]

Greptile Summary

Adds documentation for previewing and executing GDPR erasure requests against the Identities graph.

• Describes deletion scope, authentication, request and response formats, limits, and error handling.
• Links the new erasure guide from the main Identities documentation.

Confidence Score: 4/5

The invalid Snowplow ID examples should be corrected before merging because customers copying either request can fail to erase the intended identity.

Both executable examples pass a UUID as snowplow_id, contradicting the Identities schema’s required sp_-prefixed format and making the documented erasure workflow unreliable.

Files Needing Attention: docs/identities/erasure/index.md

Important Files Changed

Filename	Overview
docs/identities/erasure/index.md	Adds the complete erasure API guide, but its request and response examples use UUIDs where the documented Snowplow ID schema requires sp_-prefixed IDs.
docs/identities/index.md	Adds a valid introductory link to the new erasure guide.

_{Reviews (1): Last reviewed commit: "Add identity gdpr erasure" | Re-trigger Greptile}

[greptile-apps]

Invalid Snowplow ID format

When a customer copies either erasure request, the UUID-formatted snowplow_id contradicts the documented ^sp_[A-Za-z2-7]{26}$ schema, causing the request to target an invalid ID and fail to erase the intended identity. The corresponding response examples repeat the same invalid format.

[jamessnowplow]

Fixed in 2c65693. Both requests and both response bodies now use sp_-prefixed IDs matching the ^sp_[A-Za-z2-7]{26}$ pattern on the identity 2-0-0 entity. The truncated ... forms in the responses are gone too, so every ID on the page is copyable as-is.