feat: add index for dedup copy pipe (CM-1000)

[ulemons]

Changes

• Added minmax skipping index on updatedAt to activities.datasource, allowing ClickHouse to skip irrelevant
  granules when filtering by updatedAt
• Added minmax skipping index on updatedAt to activities_deduplicated_ds.datasource, enabling efficient range
  filtering on the watermark subquery
• Updated activities_deduplicated_copy_pipe_append_mode to replace the full table scan subquery with a
  range-limited one:

  WHERE a.updatedAt > (
      SELECT max(updatedAt)
      FROM activities_deduplicated_ds
      WHERE updatedAt > now() - INTERVAL 3 HOUR
  )
  This reduces the subquery from a full scan of 150k+ granules to only scanning recent data.
  

Known limitation

Since activities uses ReplacingMergeTree, deduplication happens in background merges. Intermediate versions of the
same activity captured between merges may result in multiple rows with the same id in activities_deduplicated_ds.
Downstream pipes consuming this datasource are expected to handle this via argMax or equivalent deduplication
patterns.

<!-- CURSOR_SUMMARY -->
---

> [!NOTE]
> **Low Risk**
> Low risk performance change: adds ClickHouse skipping indexes and narrows a watermark subquery window; primary risk is missing older late-arriving updates outside the 3-hour range.
> 
> **Overview**
> Improves performance of the `activities` → `activities_deduplicated_ds` append copy by adding `minmax` skipping indexes on `updatedAt` to both datasources.
> 
> Updates `activities_deduplicated_copy_pipe_append_mode` to compute its `updatedAt` watermark using only the last 3 hours (and clamps it with `greatest(max(updatedAt), now() - INTERVAL 3 HOUR)`), avoiding a full-table scan of `activities_deduplicated_ds` during incremental loads.
> 
> <sup>Written by [Cursor Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit c00322091bb5f27ce7edafd4c293700d6ddb50d3. This will update automatically on new commits. Configure [here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[Copilot]

Pull request overview

This PR improves incremental copying from activities into activities_deduplicated_ds by adding updatedAt-based skipping indexes and tightening the watermark subquery to a recent time window, aiming to reduce ClickHouse granule scans during scheduled COPY runs.

Changes:

• Added a minmax skipping index on updatedAt to activities.datasource.
• Added a minmax skipping index on updatedAt to activities_deduplicated_ds.datasource (and updated its sorting key).
• Updated activities_deduplicated_copy_pipe_append_mode to compute the watermark from only the last 3 hours of activities_deduplicated_ds.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
services/libs/tinybird/pipes/activities_deduplicated_copy_pipe_append_mode.pipe	Uses a range-limited watermark subquery to avoid full scans during COPY.
services/libs/tinybird/datasources/activities_deduplicated_ds.datasource	Adds updatedAt minmax skipping index; also changes sorting key to include updatedAt.
services/libs/tinybird/datasources/activities.datasource	Adds updatedAt minmax skipping index to accelerate range filters on updatedAt.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}