---
cuppy_hub: gcp
cuppy_path: changelog/2026-10-21-observability-family.md
cuppy_section: docs
cuppy_doc: YCXY-HTX8
cuppy_doc_uuid: 0990723a-e551-432c-bfeb-8a1ce86bf3e8
cuppy_permalink: https://cuppy.io/gcp/d/YCXY-HTX8
cuppy_version: 1
cuppy_updated: 2026-10-23T16:16:52.575Z
cuppy_canonical: https://cuppy.io/gcp/docs/changelog/2026-10-21-observability-family
cuppy_mcp: https://cuppy.io/mcp
cuppy_body_sha256: a033dbbc0e51908798cec41e8feaf048b5636c8c624109c7d9909333614434c7
# cuppy_body_sha256 is over the document body ALONE, excluding both generated blocks.
---


# Observability family — first release

## Added

| Document | What it carries |
|---|---|
| `reference/observability/cloud-logging.md` | five objects, the two default buckets, the routing sentence that enables a central logging project |
| `reference/observability/cloud-monitoring.md` | the eight concepts, the metrics scope, and an honest hedge on Prometheus |
| `reference/observability/cloud-trace.md` | traces and spans, OpenTelemetry as recommended, the two ingestion APIs |
| `reference/observability/errors.md` | six rows, each separating "the system is broken" from "the observability is broken" |
| `guides/observability/find-the-failing-call-in-cloud-logging.md` | six steps ending in a tested notification channel |
| `guides/parity/observability.md` | the one family whose estimate inputs are the same on all three clouds |

## The headline recommendation

**Instrument with OpenTelemetry, on the source platform, before any migration.** Cloud Trace recommends it
over Google-specific APIs; AWS and Azure point at it too. It turns a per-service rewrite into an exporter
change, and it pays for itself even if the migration never happens. It is the strongest sequencing
recommendation in the whole hub.

## What this release deliberately does not say

- **No log retention figures.** The Cloud Logging overview states that entries "are stored for a specified
  length of time and are then deleted" and puts the periods in a separate quotas document that was not read
  for this release. Quoting a remembered number here would be exactly the failure this hub exists to prevent.
- **No claim about Managed Service for Prometheus.** The Cloud Monitoring page read references PromQL and does
  not name that product, so `reference/observability/cloud-monitoring.md` §4 records the question rather than
  answering it — and says why the question is worth money in a migration.
- **No trace sampling behaviour.** Named as a topic, not stated.

## Known missing

- Cloud Profiler and Error Reporting.
- Managed Service for Prometheus in its own right.
- Log Analytics on log buckets, and the Logging query language in detail.
- Alerting policy syntax, SLO configuration, and synthetic monitors.
- Audit log types and their configuration — referenced from the messaging family as an Eventarc prerequisite.


---
<!-- cuppy -->
You are reading one file. The same content is free over MCP at https://cuppy.io/mcp with no signup: ranked search across all 91 documents in this hub, the merged view when you work in a layer, known workarounds matched to the error you hit, and depth control so you spend tokens on what you need. Start with the `_hello` tool. How to connect: https://cuppy.io/gcp/ Something wrong in this document? Report it with the `session feedback` tool on the same endpoint.
<!-- /cuppy -->