# Google Cloud hub Hub: gcp Section served: docs Start here: call the `_hello` tool on the MCP endpoint below. It says what this endpoint is and names the first document to read; every answer after it names the next call. No key and no header are needed to read a public hub. Canonical document URL: https://cuppy.io/gcp/doc/{path without .md} Raw markdown URL: https://cuppy.io/gcp/docs/{path} Raw file URL: the same path with its .md extension Sitemap: https://cuppy.io/gcp/sitemap.xml MCP endpoint: https://cuppy.io/mcp/gcp — start with the `_hello` tool: ranked search across the hub, the merged view when you work in a layer, workarounds matched to the error you hit, and depth control. Something wrong in this document? Report it with the `session feedback` tool on the same endpoint. ## Folders - changelog/ — 11 files, 10331 tokens - examples/ — 1 files, 2160 tokens - guides/ — 2 files, 4116 tokens - guides/compute/ — 1 files, 1352 tokens - guides/data/ — 1 files, 1260 tokens - guides/databases/ — 1 files, 1276 tokens - guides/iac/ — 1 files, 1393 tokens - guides/identity/ — 2 files, 2905 tokens - guides/messaging/ — 1 files, 1297 tokens - guides/migration/ — 1 files, 1739 tokens - guides/networking/ — 1 files, 1365 tokens - guides/observability/ — 1 files, 1253 tokens - guides/parity/ — 10 files, 25811 tokens - guides/storage/ — 1 files, 1321 tokens - patterns/ — 3 files, 4886 tokens - reference/compute/ — 5 files, 6761 tokens - reference/data/ — 4 files, 5333 tokens - reference/databases/ — 6 files, 8073 tokens - reference/iac/ — 4 files, 5184 tokens - reference/identity/ — 5 files, 8536 tokens - reference/messaging/ — 4 files, 5218 tokens - reference/migration/ — 5 files, 7148 tokens - reference/networking/ — 6 files, 8325 tokens - reference/observability/ — 4 files, 4960 tokens - reference/overview/ — 4 files, 5945 tokens - reference/storage/ — 5 files, 5919 tokens ## Hub pages - https://cuppy.io/gcp — the hub's front page: what it is, its folder spine with counts - https://cuppy.io/gcp/documents/{folder} — one folder, laid out as a catalog or a reading list - https://cuppy.io/gcp/doc/{path without .md} — one document, rendered - https://cuppy.io/gcp/search?q={query} — one search over this hub (on a host this hub serves directly, drop the `/gcp` prefix: /, /documents/{folder}, /doc/{path}, /search) ## Files - HUB.md — 83TH-6Q65 — Google Cloud hub — boot/markdown — 1051 tokens — asserted - changelog/2026-10-21-compute-family.md — EZJG-ST7B — Release note — compute family, 2026-10-21 — reference/markdown — 887 tokens — asserted - changelog/2026-10-21-data-family.md — 69XG-HZ89 — Release note — data and analytics family, 2026-10-21 — reference/markdown — 851 tokens — asserted - changelog/2026-10-21-databases-family.md — DGYR-MAAG — Release note — databases family, 2026-10-21 — reference/markdown — 858 tokens — asserted - changelog/2026-10-21-iac-family.md — RW3J-GMWA — Release note — infrastructure-as-code family, 2026-10-21 — reference/markdown — 736 tokens — asserted - changelog/2026-10-21-identity-family.md — XT3T-D2QA — Release note — identity family, 2026-10-21 — reference/markdown — 825 tokens — asserted - changelog/2026-10-21-messaging-family.md — YTKW-1W6N — Release note — messaging family, 2026-10-21 — reference/markdown — 733 tokens — asserted - changelog/2026-10-21-migration-family.md — PJ24-5EY0 — Release note — migration family, 2026-10-21 — reference/markdown — 934 tokens — asserted - changelog/2026-10-21-networking-family.md — KYYB-CH3R — Release note — networking family, 2026-10-21 — reference/markdown — 833 tokens — asserted - changelog/2026-10-21-observability-family.md — YCXY-HTX8 — Release note — observability family, 2026-10-21 — reference/markdown — 818 tokens — asserted - changelog/2026-10-21-release-notes.md — 1PRE-R4J9 — Release notes — Google Cloud hub, first release, 2026-10-21 — reference/markdown — 2126 tokens — asserted - changelog/2026-10-21-storage-family.md — 1EGP-F975 — Release note — storage family, 2026-10-21 — reference/markdown — 730 tokens — asserted - examples/estimate-a-three-tier-migration.md — J0P2-B8GE — Worked example — estimating a three-tier application's move onto Google Cloud — example/markdown — 2160 tokens — asserted - guides/compute/deploy-a-container-on-cloud-run.md — DE0Q-7AK2 — Deploy a container on Cloud Run without meeting the startup error — guide/markdown — 1352 tokens — asserted - guides/data/query-bigquery-from-typescript.md — DCMH-3D1K — Query BigQuery from TypeScript on Node, with the errors handled as documented — guide/markdown — 1260 tokens — asserted - guides/databases/connect-to-cloud-sql-with-iam-auth.md — 2WAR-8T70 — Connect to Cloud SQL with no database credential to manage — guide/markdown — 1276 tokens — asserted - guides/iac/terraform-first-apply.md — T9SJ-G1Q2 — A first Terraform apply against a new Google Cloud project, in the order that works — guide/markdown — 1393 tokens — asserted - guides/identity/fix-permission-denied.md — MHSM-NJWZ — Fix a permission denial on Google Cloud — the order to check things in — guide/markdown — 1523 tokens — asserted - guides/identity/github-actions-without-keys.md — 1729-764C — Authenticate GitHub Actions to Google Cloud with no service account key — guide/markdown — 1382 tokens — asserted - guides/messaging/deliver-pubsub-to-cloud-run.md — QYGE-0A5Y — Deliver Pub/Sub messages to a Cloud Run service, with the failure paths closed — guide/markdown — 1297 tokens — asserted - guides/migration/assess-before-you-move.md — SD83-S9Z6 — Assess before you move — the seven measurements that make a migration estimate defensible — guide/markdown — 1739 tokens — asserted - guides/networking/reach-a-private-service.md — ENA4-6BSK — Reach a service privately on Google Cloud, without peering the networks — guide/markdown — 1365 tokens — asserted - guides/observability/find-the-failing-call-in-cloud-logging.md — 9RNY-3QGH — Find the failing call in Cloud Logging, then alert on it — guide/markdown — 1253 tokens — asserted - guides/parity-and-migration-method.md — PMVR-NGH4 — How to use the aws, azure and gcp hubs together to estimate a migration — guide/markdown — 2741 tokens — asserted - guides/parity/compute.md — 5T4B-VDFZ — Compute parity — Compute Engine, GKE, Cloud Run and functions against EC2/EKS/Fargate/Lambda and VMs/AKS/Container Apps/Functions — guide/markdown — 2843 tokens — asserted - guides/parity/data.md — C8VN-YJS7 — Data and analytics parity — BigQuery, Dataflow and Dataproc against Redshift/EMR/Glue and Fabric/Databricks/Synapse — guide/markdown — 2273 tokens — asserted - guides/parity/databases.md — JRKM-QZVV — Database parity — Cloud SQL, AlloyDB, Spanner, Firestore and Bigtable against the AWS and Azure catalogues — guide/markdown — 2705 tokens — asserted - guides/parity/iac.md — RCDY-AANN — IaC parity — Terraform and Infra Manager against CloudFormation and ARM/Bicep, with the template ceilings side by side — guide/markdown — 2426 tokens — asserted - guides/parity/identity.md — 58DQ-2GBP — Identity parity — Google Cloud IAM against AWS IAM and Microsoft Entra / Azure RBAC, and the five differences that change an estimate — guide/markdown — 3000 tokens — asserted - guides/parity/messaging.md — RWQS-QMW5 — Messaging parity — Pub/Sub, Cloud Tasks and Eventarc against SQS/SNS/EventBridge/Kinesis and Service Bus/Event Grid/Event Hubs — guide/markdown — 2420 tokens — asserted - guides/parity/migration.md — W711-KZGC — Migration-tooling parity — Migration Center, Migrate to VMs, DMS and Storage Transfer against the AWS and Azure equivalents — guide/markdown — 2406 tokens — asserted - guides/parity/networking.md — CV2S-XTQW — Networking parity — the global VPC against regional VPCs and VNets, and the four rows that re-model an estimate — guide/markdown — 2993 tokens — asserted - guides/parity/observability.md — VNMS-FXV1 — Observability parity — Cloud Logging, Monitoring and Trace against CloudWatch/X-Ray and Azure Monitor — guide/markdown — 2204 tokens — asserted - guides/parity/storage.md — 4C9T-NWC4 — Storage parity — Cloud Storage, Persistent Disk and Filestore against S3/EBS/EFS and Blob/Disk/Files — guide/markdown — 2541 tokens — asserted - guides/storage/create-a-bucket-that-is-not-public.md — 98JC-98ZQ — Create a Cloud Storage bucket that cannot accidentally become public — guide/markdown — 1321 tokens — asserted - guides/when-the-docs-have-changed.md — 2R02-SCYM — When a Google Cloud page has changed under this hub — the drift report — guide/markdown — 1375 tokens — asserted - patterns/cloud-run-to-cloud-sql-with-iam.md — ANX6-HFVE — Pattern — a Cloud Run service reaching Cloud SQL with no database credential — pattern/markdown — 1547 tokens — asserted - patterns/github-actions-to-google-cloud-keyless.md — Q3VX-BKMN — Pattern — GitHub Actions deploys to Google Cloud with no key anywhere — pattern/markdown — 1507 tokens — asserted - patterns/lift-and-shift-a-vm-estate.md — WE5R-CT3K — Pattern — lift and shift a VM estate onto Google Cloud, with the decisions named — pattern/markdown — 1832 tokens — asserted - reference/compute/cloud-functions.md — H31B-ZPXB — Cloud Run functions and the 1st gen — the renaming, the four capability differences, and which to write today — reference/markdown — 1351 tokens — asserted - reference/compute/cloud-run.md — VR1P-T98Y — Cloud Run — services, revisions, jobs, worker pools, instances, and what a revision freezes — reference/markdown — 1418 tokens — asserted - reference/compute/compute-engine.md — X3Y5-RN6J — Compute Engine — machine families, series, the machine-type naming grammar, and where custom types are allowed — reference/markdown — 1170 tokens — asserted - reference/compute/errors.md — AN1B-14HF — Compute error catalogue — the Cloud Run startup failures, the deploy-time permission errors, 429 and 503 — reference/markdown — 1549 tokens — asserted - reference/compute/gke-cluster-modes.md — 5XHJ-Y4R7 — GKE cluster modes — Autopilot and Standard, who is responsible for what, and the billing dimension each uses — reference/markdown — 1273 tokens — asserted - reference/data/bigquery.md — S56N-6X68 — BigQuery — separated compute and storage, slots, the two compute pricing models, and the four storage billing models — reference/markdown — 1385 tokens — asserted - reference/data/dataflow.md — FQNB-BTB9 — Dataflow — Apache Beam, one model for batch and streaming, and exactly-once by default — reference/markdown — 1177 tokens — asserted - reference/data/dataproc.md — 7GTP-8NH9 — Dataproc — now documented as Managed Service for Apache Spark, three deployment shapes, and clusters that start in 90 seconds — reference/markdown — 1268 tokens — asserted - reference/data/errors.md — EZ5R-NSZ6 — Data error catalogue — the BigQuery reason strings, which are retryable, and the ones that are a cost model talking — reference/markdown — 1503 tokens — asserted - reference/databases/alloydb.md — DKS8-AA0J — AlloyDB — PostgreSQL-compatible, cluster and instance model, read pools, secondaries, and the columnar engine — reference/markdown — 1231 tokens — asserted - reference/databases/bigtable.md — Q93Y-GE4R — Bigtable — the data model, the consistency that depends on replication, and why the row key is the design — reference/markdown — 1378 tokens — asserted - reference/databases/cloud-sql.md — 530R-TFRT — Cloud SQL — the three engines, the Enterprise Plus edition, the four ways to connect, and what a maintenance update costs you — reference/markdown — 1178 tokens — asserted - reference/databases/errors.md — 9CC5-SZGT — Database error catalogue — connection refusals, the role a connector needs, consistency surprises, and the restart you scheduled — reference/markdown — 1761 tokens — asserted - reference/databases/firestore.md — 8HRA-YX5S — Firestore — the two database modes, what each is for, and the MongoDB-compatibility note — reference/markdown — 1203 tokens — asserted - reference/databases/spanner.md — W1YW-2365 — Spanner — the three instance configurations, their stated availability, and the write-latency trade they encode — reference/markdown — 1322 tokens — asserted - reference/iac/config-connector.md — BTN1-5S6S — Config Connector — Google Cloud resources as Kubernetes objects, the identity that reconciles them, and the two modes — reference/markdown — 1163 tokens — asserted - reference/iac/deployment-manager-status.md — XD6V-NSWD — Cloud Deployment Manager — end of support, and the two named successors — reference/markdown — 998 tokens — asserted - reference/iac/errors.md — YWX9-4X92 — IaC error catalogue — the API that is not enabled, the reconciler's identity, state locking, and a retired product — reference/markdown — 1573 tokens — asserted - reference/iac/terraform-on-google-cloud.md — SKG2-X38W — Terraform on Google Cloud — the gcs backend, blueprints, Infrastructure Manager, and what the vendor does not prescribe — reference/markdown — 1450 tokens — asserted - reference/identity/errors.md — HBJ7-0Z5T — Identity error catalogue — access denied, actAs, quota refusals, and the one tool that answers "why" — reference/markdown — 1888 tokens — asserted - reference/identity/iam-model.md — BAQP-PGEN — Google Cloud IAM — principals, roles, permissions, allow and deny policies, inheritance, eventual consistency — reference/markdown — 1594 tokens — asserted - reference/identity/predefined-roles-and-permissions.md — BCVT-66E9 — Predefined roles and permissions — the naming format, worked role names, custom-role limits, service accounts as identity and as resource — reference/markdown — 1586 tokens — asserted - reference/identity/secret-manager-and-kms.md — TZT4-NMR5 — Secret Manager and Cloud KMS — object models, the accessor role, replication, protection levels, and what cannot be undone — reference/markdown — 1824 tokens — asserted - reference/identity/workload-identity-federation.md — Y7CS-792Z — Workload Identity Federation — pools, providers, principal strings, direct resource access, and the GKE form — reference/markdown — 1644 tokens — asserted - reference/messaging/cloud-tasks.md — PTE2-E006 — Cloud Tasks — queues of individually-addressed work, with rate limits, retries and a scheduled time — reference/markdown — 1151 tokens — asserted - reference/messaging/errors.md — MK0K-HF95 — Messaging error catalogue — redelivery that is not a bug, push endpoints that refuse, dead-lettering, and the trigger that never fires — reference/markdown — 1527 tokens — asserted - reference/messaging/eventarc.md — 513Q-8Q8R — Eventarc — CloudEvents from Google Cloud itself, the two editions, and the destinations it can reach — reference/markdown — 1277 tokens — asserted - reference/messaging/pubsub.md — 1FBQ-89SG — Pub/Sub — topics, subscriptions, the five subscription types, and the delivery guarantee you have to ask for — reference/markdown — 1263 tokens — asserted - reference/migration/database-migration-service.md — KS71-6KKH — Database Migration Service — the four homogeneous paths, the four heterogeneous ones, and what "minimal downtime" rests on — reference/markdown — 1375 tokens — asserted - reference/migration/errors.md — NVNF-NXTJ — Migration error catalogue — the failures that are permissions, networking, capacity or a version you did not check — reference/markdown — 1713 tokens — asserted - reference/migration/migrate-to-virtual-machines.md — QQ7K-1M6W — Migrate to Virtual Machines — the supported sources including AWS and Azure, and the four-project object model — reference/markdown — 1329 tokens — asserted - reference/migration/migration-center.md — CCZQ-TM58 — Migration Center — the assessment platform, its three collection methods, and the six outputs an estimate can actually use — reference/markdown — 1379 tokens — asserted - reference/migration/storage-transfer-service.md — WZ6Z-Y1K6 — Storage Transfer Service — S3, Azure Blob, HDFS and POSIX sources, scheduled or event-driven, and the 1 TiB threshold — reference/markdown — 1352 tokens — asserted - reference/networking/cloud-armor.md — 0R8V-400D — Cloud Armor — which load balancers it attaches to, the rules language, the OWASP CRS versions, and what it protects against — reference/markdown — 1315 tokens — asserted - reference/networking/cloud-load-balancing.md — E2QX-4BPM — Cloud Load Balancing — the current type names, which are global, and proxy against passthrough — reference/markdown — 1381 tokens — asserted - reference/networking/cloud-nat.md — 3QD2-VQ3H — Cloud NAT — a software-defined feature of the network, not an appliance, and it never accepts an inbound connection — reference/markdown — 1219 tokens — asserted - reference/networking/errors.md — QA71-2AQP — Networking error catalogue — the failures that present as a timeout, and how to tell them apart — reference/markdown — 1627 tokens — asserted - reference/networking/private-service-connect.md — RM7V-E873 — Private Service Connect — five object types, which side owns each, and why it beats peering for service access — reference/markdown — 1453 tokens — asserted - reference/networking/vpc.md — NRQV-H34Y — VPC — a global network with regional subnets, the two subnet modes, the implied firewall rules, and non-transitive peering — reference/markdown — 1330 tokens — asserted - reference/observability/cloud-logging.md — 2C6W-05H7 — Cloud Logging — buckets, sinks, views, log-based metrics, and the two buckets every project already has — reference/markdown — 1274 tokens — asserted - reference/observability/cloud-monitoring.md — MTGK-6SGE — Cloud Monitoring — the eight concepts, the metrics scope, and where PromQL fits — reference/markdown — 1264 tokens — asserted - reference/observability/cloud-trace.md — JZDZ-SFTM — Cloud Trace — distributed tracing with OpenTelemetry as the recommended instrumentation — reference/markdown — 1018 tokens — asserted - reference/observability/errors.md — 02YH-299H — Observability error catalogue — the empty chart, the alert that never fires, the missing log, and the trace that was not sampled — reference/markdown — 1404 tokens — asserted - reference/overview/projects-and-resource-hierarchy.md — Z6KA-5JJ1 — Projects and the resource hierarchy — organization, folders, projects, and the three project identifiers — reference/markdown — 1302 tokens — asserted - reference/overview/quotas-and-limits-model.md — DTYH-EFYF — Quotas and system limits — the two quota types, the scopes, the roles that read them, and the calls that return a live value — reference/markdown — 1493 tokens — asserted - reference/overview/regions-zones-and-locations.md — B0AG-XVR3 — Regions, zones and location scope — zonal, regional and global resources, and what that costs on a migration — reference/markdown — 1232 tokens — asserted - reference/overview/services-and-identities.md — Y4AK-S1YE — The map — every service this hub covers, where its documents are, and which identity makes the call — reference/markdown — 1918 tokens — asserted - reference/storage/cloud-storage-access-control.md — G76G-EQ9R — Cloud Storage access control — uniform bucket-level access, the 90-day window, public access prevention, and the five mechanisms — reference/markdown — 1204 tokens — asserted - reference/storage/cloud-storage.md — 5EJY-KT49 — Cloud Storage — storage classes, minimum storage durations, location types, and what changing a class does not do — reference/markdown — 1161 tokens — asserted - reference/storage/errors.md — 4J28-3C7T — Storage error catalogue — the Cloud Storage JSON API reason strings, with cause, check and fix — reference/markdown — 1416 tokens — asserted - reference/storage/filestore.md — XD6A-2DF2 — Filestore — the service tiers, their protocols, their capacity ranges, and the legacy tiers to avoid — reference/markdown — 902 tokens — asserted - reference/storage/persistent-disk-and-hyperdisk.md — WGFX-1ZR0 — Block storage — Hyperdisk, Persistent Disk, Local SSD, and the one that loses your data on a stop — reference/markdown — 1236 tokens — asserted files_listed: 91 / files_total: 91