fix: gerrit repositories filtering (CM-1079)

[joanagmaia]

Problem

Since the bucketing architecture was introduced, the Active Contributors widget (and any widget backed by activities_filtered) showed no data for projects with Gerrit integrations, even when activity data existed.

Root cause: The enrichment/cleaning copy pipes (activityRelations_bucket_clean_enrich_copy_pipe_X) validate git-platform activities by checking:
(channel, segmentId) IN (SELECT r.url, r.segmentId FROM repositories ...)

Gerrit activities store their channel in /q/project: format (e.g. https://gerrit.example.com/r/q/project:myproject), but repositories.url stores the base URL (https://gerrit.example.com/r/myproject). The match always fails, so all Gerrit activities were silently dropped during the cleaning step and never made it into the cleaned bucket datasources that widgets query.

The repos_to_channels pipe already handled this URL expansion correctly at query time — but the cleaning step wasn't using it.

Fix

Extended repos_to_channels.pipe to also output segmentId alongside channel. This is backward-compatible: all 21 existing consumers only SELECT channel and are unaffected.

The 10 cleaning pipes now delegate to repos_to_channels instead of maintaining an inline subquery. This means the Gerrit URL expansion logic lives in exactly one place — any future change to channel formats only needs updating in repos_to_channels.pipe.

---

Note

Medium Risk
Touches the activity cleaning/enrichment COPY pipelines that populate downstream analytics datasets, so mistakes could silently drop or include activity data. Logic change is localized but affects all bucket shards and repo/channel matching semantics.

Overview
Fixes Gerrit repo activities being filtered out during the activityRelations_bucket_clean_enrich_copy_pipe_* cleaning step by replacing the inline repositories/insightsProjects subquery with (channel, segmentId) IN (SELECT channel, segmentId FROM repos_to_channels).

Extends repos_to_channels.pipe to also output segmentId alongside each expanded channel (including Gerrit /q/project: variants) and tightens selection to enabled, non-deleted repos/projects, centralizing repo-to-activity-channel matching in one place.

^{Written by Cursor Bugbot for commit 0c39d89. This will update automatically on new commits. Configure here.}

[Copilot]

Pull request overview

Fixes Gerrit-backed projects showing no data in widgets that rely on activities_filtered by ensuring the bucket cleaning/enrichment COPY pipes validate git-platform activities against the expanded set of possible repository channel formats (including Gerrit /q/project: variants).

Changes:

• Extend repos_to_channels.pipe to also return segmentId alongside channel and to filter out repos belonging to deleted insightsProjects.
• Update the 10 activityRelations_bucket_clean_enrich_copy_pipe_{0..9}.pipe cleaning pipes to validate (channel, segmentId) via repos_to_channels instead of an inline repositories subquery.

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
services/libs/tinybird/pipes/repos_to_channels.pipe	Adds segmentId output and keeps Gerrit channel expansion centralized for reuse by cleaning/validation logic.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_0.pipe	Switches repo validation subquery to repos_to_channels to include Gerrit channel variants.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_1.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_2.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_3.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_4.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_5.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_6.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_7.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_8.pipe	Same change as bucket 0.
services/libs/tinybird/pipes/activityRelations_bucket_clean_enrich_copy_pipe_9.pipe	Same change as bucket 0.

Comments suppressed due to low confidence (1)

services/libs/tinybird/pipes/repos_to_channels.pipe:44

• When repos is provided, repos_to_expand hard-codes segmentId to '' (lines 19-24), but gerrit_repos later rehydrates segmentId from repositories (lines 40-44). This makes the output inconsistent: non-Gerrit rows will have empty segmentId, while Gerrit variants can have a real segmentId, which also contradicts the response description. Consider making segmentId consistently empty for all outputs when repos is provided (e.g., carry segmentId from repos_to_expand into Gerrit variants), or update the contract/documentation and ensure all branches follow it.

    {% if defined(repos) %}
        SELECT
            arrayJoin(
                {{ Array(repos, 'String', description="Repository URLs to expand", required=False) }}
            ) AS url,
            '' AS segmentId
    {% else %}
        SELECT r.url, r.segmentId
        FROM repositories r FINAL
        INNER JOIN insightsProjects i FINAL ON r.insightsProjectId = i.id
        WHERE
            isNull (r.deletedAt) AND r.enabled = true AND isNull (i.deletedAt)
            {% if defined(excluded) and excluded %} AND r.excluded = true
            {% end %}
    {% end %}

NODE gerrit_repos
DESCRIPTION >
    Identify Gerrit repositories by joining with integrations table

SQL >
    SELECT r.url, r.segmentId
    FROM repositories r FINAL
    JOIN integrations i FINAL ON r.sourceIntegrationId = i.id
    WHERE i.platform = 'gerrit' AND isNull (r.deletedAt) AND r.url IN (SELECT url FROM repos_to_expand)

---

💡 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.}

[mbani01]

LGTM