chore(DENG-11021): complete certain tier 2 backfills

[kbammarito]

Description

Completes the two Tier 2 cohort_daily_statistics install_source backfills initiated in #9690:

• telemetry_derived.cohort_daily_statistics_v2
• glean_telemetry_derived.cohort_daily_statistics_v1

Validated both staging tables. Reading one in-range activity_date per check gives complete cohort coverage (the query LEFT JOINs actives onto the full in-range cohort, so every cohort emits all its buckets at full size each day).

Within-staging structural gates — clean for both families:

• No duplicate rows on the full dimension key now that install_source is part of it.
• Grain strictly got finer — the install_source split happened, not a no-op.
• Pre-2025 cohorts stay install_source = NULL at one row per dimension combo (no bleed-through, no fan-out).
• Non-fenix rows are all install_source = NULL.

Live-prod comparisons — drift-tolerant (frozen snapshot gone; prod shredder-shrunk since build), both benign:

• Cohort-size reconciliation: uniform +0.9% for telemetry (staging >= prod, no recent-cohort spike) — expected shredder gap.
• Value existence: 112 telemetry orphan pairs, all 1-2-client buckets with real rare installer package names (shredder-emptied from prod, not invented).
• Glean: exactly clean, 0 rows on every check.

Both families promote-safe.

Related Tickets & Documents

• DENG-11021
• DENG-11362

Reviewer, please follow this checklist

[github-actions]

This PR flips the status field from Initiate to Complete on the two cohort_daily_statistics backfill entries (telemetry_derived.cohort_daily_statistics_v2 and glean_telemetry_derived.cohort_daily_statistics_v1) added on 2026-07-06 for the install_source change from #9690. This is the documented transition (docs/cookbooks/backfilling_a_table.md) that triggers the staging→production swap after staging validation.

I reviewed both files against the backfill schema and validation rules and have no inline findings — the change is consistent with the standard workflow, and the accompanying validation notes in the description cover the staging/prod reconciliation.

[scholtzan]

Integration report

• View full diff, if available (SQL + DAGs) (requires private-bigquery-etl access)
• Private CI run
• Bigeye monitoring dry-run: ✅ plan clean — view plan (requires private-bigquery-etl access)