Skip to content

Update warehouse docs - #1908

Open
Agnes Kiss (agnes-kiss) wants to merge 2 commits into
mainfrom
update-signals-sync
Open

Update warehouse docs#1908
Agnes Kiss (agnes-kiss) wants to merge 2 commits into
mainfrom
update-signals-sync

Conversation

@agnes-kiss

Copy link
Copy Markdown
Contributor

What changed?

Refined warehouse / batch engine + SDK related signals docs as per the upcoming release (and anything recent that we have not yet updated).

@agnes-kiss Agnes Kiss (agnes-kiss) added do not merge Flag to denote a Issue or PR which should not yet be merged (usually pending a release) and removed cla:yes labels Aug 13, 2026
@claude

claude Bot commented Aug 13, 2026

Copy link
Copy Markdown

Documentation style review

Scoped to writing quality, style guide compliance, structure, and links in the diff. The Databricks support updates and the new image (descriptive filename in images/, good alt text) are clean. The findings below are mostly in the new Console-tab prose in warehouse-config/index.md and the new admonition in concepts/index.md.

1. Grammar error in the new admonition

docs/signals/concepts/index.md:121

"When backfilling, the events table filtered on the load_tstamp column, since this is typically the partition or cluster key on Snowplow-loaded event tables (by default, or per deployment's instructions)."

Missing verb ("the events table is filtered"), and the sentence is passive where the style guide asks for active voice ("Use active voice: 'the Collector receives events'"). The parenthetical "(by default, or per deployment's instructions)" is vague enough that it doesn't add information.

Suggested fix:

Signals filters the events table on the `load_tstamp` column. This column is typically the partition or cluster key on Snowplow-loaded event tables, so filtering on it lets queries prune partitions efficiently.

2. Admonitions are missing bracket titles

Style guide, Markdown formatting: "Always set custom headings using bracket syntax: :::note[Title]." Three new admonitions have none:

  • concepts/index.md:120:::info
  • warehouse-config/index.md:45:::warning
  • warehouse-config/index.md:87:::warning

For example, :::warning[Use a last-modified timestamp] and :::info[Backfill query performance].

3. Backticks used for Console UI labels instead of bold

CLAUDE.md: "Use bold only for UI elements (buttons, page titles)" and "Use backticks for code, file names, and table/field names." The new Console tab uses backticks for form fields:

"Select the attribute key you have data for" … "setting the TTL" … "the sync frequency selector" … "specify a backfill start date" … "Select which table to sync from" … "select which table fields you want to send to Signals"

These are UI labels, so they should be bold and match the label casing shown in Console: Attribute key, TTL, Sync frequency, Backfill start date, Table, Table fields. The existing line just above them already does this correctly: "select Warehouse as the data source."

Related inconsistency in the same paragraph (line 43): `Modified date` is backticked twice but written plain once ("The Modified date column must be in UTC"). Pick one treatment — bold, since it's the UI label.

4. "currently" violates the present-tense rule

Style guide: "Stay in the present tense. Never use 'currently', 'now', 'in the future', 'latest'." Three touched lines carry it over:

  • attributes/attribute-groups/index.md:92
  • get-started/index.md:34
  • setup/index.md:23

"Snowflake, BigQuery, and Databricks are supported currently."

Since you're already editing these lines, drop the word: "Snowflake, BigQuery, and Databricks are supported." (Applies to all three occurrences.)

5. Tense and wording in the sync frequency paragraph

warehouse-config/index.md:31-33

"Decide on how long you would like the data to persist in the profile store by setting the TTL."

"By default, the updates will run on an hourly basis, which you can adjust to a lower cadence by the sync frequency selector (6h, 12h, 24h)."

Three issues: future tense ("will run") where the style guide asks for present tense; wordiness ("Decide on how long you would like", "run on an hourly basis"); and "lower cadence" is ambiguous — 6h/12h/24h is less frequent, which reads as a higher number.

Suggested fix:

Set the **TTL** to control how long the data persists in the Profiles Store.

Updates run hourly by default. Use the **Sync frequency** selector to sync less often (6h, 12h, or 24h).

6. "profile store" should be "Profiles Store"

warehouse-config/index.md:31 is the only place in docs/signals that writes "profile store" lowercase and singular. Every other reference, including line 14 of this same file, uses "Profiles Store".

7. Comma splice in the SDK tab

warehouse-config/index.md:83

"timestamp_field must be in UTC, a column in local time will cause rows to be silently matched against the wrong sync window, with no error raised."

Two independent clauses joined by a comma. Split into two sentences: "timestamp_field must be in UTC. A column in local time causes rows to be silently matched against the wrong sync window, with no error raised." (Also drops a "will" for present tense.)

8. Repeated "avoid" in the Console warning

warehouse-config/index.md:46

"Avoid using a business or event timestamp that doesn't change when a row is reloaded or updated to avoid data being missed."

"Avoid … to avoid" reads awkwardly, and this version omits the reason the SDK-tab warning gives. Align the two tabs on the clearer SDK wording:

Don't use a business or event timestamp that doesn't change when a row is reloaded or updated — Signals won't be able to tell the row has changed, and the data may be missed.

9. The page no longer has any H2 sections

Removing ## Provide source table details and ## Define which fields to sync leaves warehouse-config/index.md with only H3s, all of them inside <TabItem> blocks. CLAUDE.md treats H2 as the main section level ("H2 is for main sections. H3 is for subsections"), so H3s with no parent H2 leave the page with an empty table of contents and an implicit level skip.

If the intent is to move the steps inside the tabs, consider keeping a wrapping H2 above <Tabs> (for example ## Configure a warehouse source) with the existing intro prose under it, then the H3s inside each tab as subsections.

Minor, same file: there's a stray double blank line at lines 90-91 before ### Define which fields to sync.

@cloudflare-workers-and-pages

cloudflare-workers-and-pages Bot commented Aug 13, 2026

Copy link
Copy Markdown

Deploying with  Cloudflare Workers  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 4bb8fc6 Commit Preview URL

Branch Preview URL
Aug 13 2026, 06:05 PM

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

do not merge Flag to denote a Issue or PR which should not yet be merged (usually pending a release)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants