## TL;DR
Enterprise accounts going dark together is rarely their doing. They share the special paths: allowlisted ingestion, self-hosted or VPC deployments, stricter consent handling, and region-pinned data. One change to any shared enterprise path blinds exactly that segment while self-serve data keeps flowing.

```text
usage data missing for enterprise accounts: telemetry pipeline broken
```

## Use this when
- Usage is missing for many or all enterprise accounts at once
- Self-serve and mid-market data is normal
- An infrastructure, region, or allowlist change happened recently

## Not for this skill when
- One enterprise customer changed their own deployment
- Enterprise contracts were paused or terminated

## Steps
1. List affected accounts and what they share: deployment type, region, ingestion route, and contract tier. Verify: the shared attribute that defines the blind segment is found
2. Test ingestion end to end on that shared path with a synthetic account. Verify: the synthetic event either lands or dies at a named hop
3. Review recent changes to allowlists, region routing, and enterprise gateway config in the failure window. Verify: the change that excluded the segment is identified with its timestamp
4. Restore the path, replay queued or buffered enterprise events where the source retained them. Verify: enterprise usage reappears and the replay fills the gap without duplicates
5. Add per-segment ingestion volume alerts so one tier going dark pages separately from global volume. Verify: a future enterprise-only outage alerts within the hour

## Variant phrasings
### enterprise accounts have no usage data
A related phrasing for the same underlying problem.
### telemetry missing for largest customers
A related phrasing for the same underlying problem.
### enterprise segment dark in analytics
A related phrasing for the same underlying problem.

Compatibility: Compatibility: segmented ingestion pipelines with VPC or region routing and warehouse rollups, current versions.

## Why it happens
Segment-shaped outages come from segment-shaped configuration. Enterprise traffic deliberately takes different roads, so a global health check stays green while the enterprise road is closed. Testing the exact road those accounts travel is the only reliable check.

## Edge cases / pitfalls
- Some enterprise gaps are contractual: data-processing terms may pause telemetry during renegotiation. Verify with the account team before replaying.
- Buffered events at the edge expire. Note the retention limit before promising a full backfill.
- Region routing fixes can move data between warehouses. Confirm which store reports read before declaring recovery.

## Provenance

Resolved from the public thread: https://vectle.com/posts/pst_j-Q76M9xeYlrJwKox2IO5A