dt-obs-gcp

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

GCP Cloud Infrastructure

GCP云基础设施

Monitor and analyze GCP resources using Dynatrace Smartscape and DQL. Query GCP services, manage organizational hierarchy, audit security posture, and track resource ownership across your GCP infrastructure.

使用Dynatrace Smartscape和DQL监控并分析GCP资源。查询GCP服务、管理组织层级、审计安全态势，并追踪整个GCP基础设施中的资源归属。

When to Use This Skill

何时使用该技能

Use this skill when the user needs to work with GCP resources in Dynatrace. Load the reference file for the task type:

Task	File to load
Inventory and topology queries	(no additional file — use core patterns above)
Compute Engine instances, machine types, IP addresses	Load `references/compute-instances.md`
GKE clusters, node pools, pods, deployments, services, RBAC	Load `references/kubernetes-gke.md`
Cloud Run services, revisions, executions	Load `references/serverless-containers.md`
VPC networks, subnets, routes, DNS records	Load `references/networking-dns.md`
Pub/Sub topics	Load `references/messaging-pubsub.md`
IAM service accounts, roles, Secret Manager	Load `references/iam-security.md`
Monitoring dashboards, logging, saved queries	Load `references/monitoring-logging.md`
GCP projects, regions, organizational hierarchy	Load `references/resource-management.md`
Resource ownership, GCP labels, organizational structure	Load `references/resource-ownership.md`

当用户需要在Dynatrace中处理GCP资源时使用本技能。根据任务类型加载参考文件：

任务	需加载的文件
资源清单与拓扑查询	（无需额外文件 — 使用上方核心模式）
Compute Engine实例、机器类型、IP地址	加载 `references/compute-instances.md`
GKE集群、节点池、Pod、部署、服务、RBAC	加载 `references/kubernetes-gke.md`
Cloud Run服务、修订版本、执行记录	加载 `references/serverless-containers.md`
VPC网络、子网、路由、DNS记录	加载 `references/networking-dns.md`
Pub/Sub主题	加载 `references/messaging-pubsub.md`
IAM服务账号、角色、Secret Manager	加载 `references/iam-security.md`
监控仪表板、日志、已保存查询	加载 `references/monitoring-logging.md`
GCP项目、区域、组织层级	加载 `references/resource-management.md`
资源归属、GCP标签、组织结构	加载 `references/resource-ownership.md`

Core Concepts

核心概念

Entity Types

实体类型

GCP resources use the

GCP_*

prefix and can be queried using the

smartscapeNodes

function. All GCP entities are automatically discovered and modeled in Dynatrace Smartscape.

Compute:

GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE

GCP_COMPUTE_GOOGLEAPIS_COM_ADDRESS

Networking:

GCP_COMPUTE_GOOGLEAPIS_COM_NETWORK

GCP_COMPUTE_GOOGLEAPIS_COM_SUBNETWORK

GCP_COMPUTE_GOOGLEAPIS_COM_ROUTE

GCP_DNS_GOOGLEAPIS_COM_RESOURCERECORDSET

Kubernetes (GKE):

GCP_K8S_IO_POD

GCP_K8S_IO_NODE

GCP_K8S_IO_SERVICE

GCP_K8S_IO_SERVICEACCOUNT

GCP_K8S_IO_PERSISTENTVOLUMECLAIM

GCP_APPS_K8S_IO_DEPLOYMENT

GCP_APPS_K8S_IO_STATEFULSET

GCP_CONTAINER_GOOGLEAPIS_COM_NODEPOOL

GCP_RBAC_AUTHORIZATION_K8S_IO_CLUSTERROLEBINDING

GCP_RBAC_AUTHORIZATION_K8S_IO_ROLEBINDING

Serverless:

GCP_RUN_GOOGLEAPIS_COM_SERVICE

GCP_RUN_GOOGLEAPIS_COM_REVISION

GCP_RUN_GOOGLEAPIS_COM_EXECUTION

IAM & Security:

GCP_IAM_GOOGLEAPIS_COM_SERVICEACCOUNT

GCP_IAM_GOOGLEAPIS_COM_ROLE

GCP_SECRETMANAGER_GOOGLEAPIS_COM_SECRETVERSION

Messaging:

GCP_PUBSUB_GOOGLEAPIS_COM_TOPIC

Monitoring:

GCP_MONITORING_GOOGLEAPIS_COM_DASHBOARD

GCP_LOGGING_GOOGLEAPIS_COM_SAVEDQUERY

Infrastructure:

GCP_REGION

GCP资源使用

GCP_*

前缀，可通过

smartscapeNodes

函数查询。所有GCP实体都会被自动发现并在Dynatrace Smartscape中建模。

计算类：

GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE

、

GCP_COMPUTE_GOOGLEAPIS_COM_ADDRESS

网络类：

GCP_COMPUTE_GOOGLEAPIS_COM_NETWORK

、

GCP_COMPUTE_GOOGLEAPIS_COM_SUBNETWORK

、

GCP_COMPUTE_GOOGLEAPIS_COM_ROUTE

、

GCP_DNS_GOOGLEAPIS_COM_RESOURCERECORDSET

Kubernetes（GKE）：

GCP_K8S_IO_POD

、

GCP_K8S_IO_NODE

、

GCP_K8S_IO_SERVICE

、

GCP_K8S_IO_SERVICEACCOUNT

、

GCP_K8S_IO_PERSISTENTVOLUMECLAIM

、

GCP_APPS_K8S_IO_DEPLOYMENT

、

GCP_APPS_K8S_IO_STATEFULSET

、

GCP_CONTAINER_GOOGLEAPIS_COM_NODEPOOL

、

GCP_RBAC_AUTHORIZATION_K8S_IO_CLUSTERROLEBINDING

、

GCP_RBAC_AUTHORIZATION_K8S_IO_ROLEBINDING

无服务器类：

GCP_RUN_GOOGLEAPIS_COM_SERVICE

、

GCP_RUN_GOOGLEAPIS_COM_REVISION

、

GCP_RUN_GOOGLEAPIS_COM_EXECUTION

IAM与安全类：

GCP_IAM_GOOGLEAPIS_COM_SERVICEACCOUNT

、

GCP_IAM_GOOGLEAPIS_COM_ROLE

、

GCP_SECRETMANAGER_GOOGLEAPIS_COM_SECRETVERSION

消息类：

GCP_PUBSUB_GOOGLEAPIS_COM_TOPIC

监控类：

GCP_MONITORING_GOOGLEAPIS_COM_DASHBOARD

、

GCP_LOGGING_GOOGLEAPIS_COM_SAVEDQUERY

基础设施类：

GCP_REGION

Common GCP Fields

通用GCP字段

All GCP entities include:

```
gcp.project.id
```
— GCP project identifier
```
gcp.region
```
— GCP region (e.g., us-central1)
```
gcp.zone
```
— GCP zone (e.g., us-central1-a)
```
gcp.organization.id
```
— GCP organization identifier
```
gcp.resource.name
```
— Resource name
```
gcp.resource.type
```
— Resource type identifier
```
gcp.asset.type
```
— GCP asset type
```
gcp.object
```
— JSON blob containing full resource configuration

所有GCP实体均包含以下字段：

```
gcp.project.id
```
— GCP项目标识符
```
gcp.region
```
— GCP区域（例如：us-central1）
```
gcp.zone
```
— GCP可用区（例如：us-central1-a）
```
gcp.organization.id
```
— GCP组织标识符
```
gcp.resource.name
```
— 资源名称
```
gcp.resource.type
```
— 资源类型标识符
```
gcp.asset.type
```
— GCP资产类型
```
gcp.object
```
— 包含完整资源配置的JSON二进制大对象

GCP Organizational Hierarchy

GCP组织层级

GCP resources are organized in a hierarchy:

Organization — Top-level container (
```
gcp.organization.id
```
)
Folder — Logical grouping within an organization
Project — Resource container (
```
gcp.project.id
```
)
Region/Zone — Physical location (
```
gcp.region
```
,
```
gcp.zone
```
)

GCP资源按以下层级组织：

组织 — 顶层容器（
```
gcp.organization.id
```
）
文件夹 — 组织内的逻辑分组
项目 — 资源容器（
```
gcp.project.id
```
）
区域/可用区 — 物理位置（
```
gcp.region
```
、
```
gcp.zone
```
）

Entity Naming Convention

实体命名规范

GCP entity types follow the pattern

GCP_<SERVICE_API>_<RESOURCE>

Service API maps to the Google API domain (e.g.,
```
compute.googleapis.com
```
→
```
COMPUTE_GOOGLEAPIS_COM
```
)
Resource is the specific resource type (e.g.,
```
INSTANCE
```
,
```
NETWORK
```
)

Examples:

```
GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE
```
— Compute Engine VM
```
GCP_K8S_IO_POD
```
— GKE pod
```
GCP_RUN_GOOGLEAPIS_COM_SERVICE
```
— Cloud Run service

GCP实体类型遵循

GCP_<SERVICE_API>_<RESOURCE>

模式：

服务API对应Google API域名（例如：
```
compute.googleapis.com
```
→
```
COMPUTE_GOOGLEAPIS_COM
```
）
资源指具体的资源类型（例如：
```
INSTANCE
```
、
```
NETWORK
```
）

示例：

```
GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE
```
— Compute Engine虚拟机
```
GCP_K8S_IO_POD
```
— GKE Pod
```
GCP_RUN_GOOGLEAPIS_COM_SERVICE
```
— Cloud Run服务

Query Patterns

查询模式

All GCP queries build on four core patterns. Master these and adapt them to any entity type.

所有GCP查询都基于四种核心模式。掌握这些模式后可适配至任意实体类型。

Pattern 1: Resource Discovery

模式1：资源发现

List resources by type, filter by project/region/zone, summarize counts:

dql

smartscapeNodes "GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE"
| fields name, gcp.project.id, gcp.region, gcp.zone, gcp.resource.name

To list all GCP resource types, replace with

"GCP_*"

and add

| summarize count = count(), by: {type} | sort count desc

. Add filters like

| filter gcp.project.id == "<PROJECT_ID>"

| filter gcp.region == "<REGION>"

to scope results.

按类型列出资源，按项目/区域/可用区筛选，汇总数量：

dql

smartscapeNodes "GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE"
| fields name, gcp.project.id, gcp.region, gcp.zone, gcp.resource.name

若要列出所有GCP资源类型，将上述语句替换为

"GCP_*"

并添加

| summarize count = count(), by: {type} | sort count desc

。可添加

| filter gcp.project.id == "<PROJECT_ID>"

或

| filter gcp.region == "<REGION>"

等筛选条件来限定结果范围。

Pattern 2: Configuration Parsing

模式2：配置解析

Parse

gcp.object

JSON for detailed configuration fields:

dql

smartscapeNodes "GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE"
| parse gcp.object, "JSON:gcpjson"
| fieldsAdd machineType = gcpjson[configuration][resource][machineType],
            status = gcpjson[configuration][resource][status]
| fields name, gcp.project.id, machineType, status

GCP configuration fields are nested under

gcpjson[configuration][resource][...]

for primary resource attributes and

gcpjson[configuration][additionalAttributes][...]

for extended properties.

解析

gcp.object

JSON以获取详细配置字段：

dql

smartscapeNodes "GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE"
| parse gcp.object, "JSON:gcpjson"
| fieldsAdd machineType = gcpjson[configuration][resource][machineType],
            status = gcpjson[configuration][resource][status]
| fields name, gcp.project.id, machineType, status

GCP配置字段中，主要资源属性嵌套在

gcpjson[configuration][resource][...]

下，扩展属性嵌套在

gcpjson[configuration][additionalAttributes][...]

下。

Pattern 3: Relationship Traversal

模式3：关系遍历

Follow relationships between resources:

dql

smartscapeNodes "GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE"
| traverse "*", "GCP_COMPUTE_GOOGLEAPIS_COM_SUBNETWORK"
| fields name, gcp.project.id

GCP entities use

"*"

as the relationship name in traversals because GCP entities do not have named relationship types. Use

fieldsKeep

to carry fields through traversals and

dt.traverse.history[-N]

to access ancestor fields.

追踪资源间的关系：

dql

smartscapeNodes "GCP_COMPUTE_GOOGLEAPIS_COM_INSTANCE"
| traverse "*", "GCP_COMPUTE_GOOGLEAPIS_COM_SUBNETWORK"
| fields name, gcp.project.id

GCP实体在遍历时使用

"*"

作为关系名称，因为GCP实体没有命名的关系类型。使用

fieldsKeep

在遍历过程中保留字段，使用

dt.traverse.history[-N]

访问祖先字段。

Pattern 4: Label-Based Ownership

模式4：基于标签的归属管理

Group resources by GCP labels for ownership and organizational tracking:

dql

smartscapeNodes "GCP_*"
| filter isNotNull(`tags:gcp_labels`)
| fields name, gcp.project.id, `tags:gcp_labels`

GCP labels are exposed via the

tags:gcp_labels

field and must be accessed using backtick syntax. Replace

"GCP_*"

with a specific type to scope to one service.

按GCP标签对资源进行分组，用于归属和组织追踪：

dql

smartscapeNodes "GCP_*"
| filter isNotNull(`tags:gcp_labels`)
| fields name, gcp.project.id, `tags:gcp_labels`

GCP标签通过

tags:gcp_labels

字段暴露，必须使用反引号语法访问。将

"GCP_*"

替换为特定类型可限定至单个服务。

Reference Guide

参考指南

Load reference files for detailed queries when the core patterns above need service-specific adaptation.

Reference	When to load	Key content
compute-instances.md	Compute Engine VMs, machine types, IP addresses, disks	Instance inventory, machine type distribution, status checks
kubernetes-gke.md	GKE clusters, node pools, pods, deployments, services, RBAC	Cluster topology, workload distribution, RBAC bindings
serverless-containers.md	Cloud Run services, revisions, executions	Service inventory, revision tracking, execution analysis
networking-dns.md	VPC networks, subnets, routes, DNS records	Network topology, subnet analysis, route tables, DNS record sets
messaging-pubsub.md	Pub/Sub topics	Topic inventory, messaging topology
iam-security.md	IAM service accounts, roles, Secret Manager	Service account audit, role analysis, secret version tracking
monitoring-logging.md	Monitoring dashboards, logging, saved queries	Dashboard inventory, saved query analysis
resource-management.md	GCP projects, regions, organizational hierarchy	Project inventory, region distribution, hierarchy mapping
resource-ownership.md	Resource ownership, GCP labels, organizational structure	Label-based grouping, project-level summaries, chargeback

当核心模式需要针对特定服务调整时，加载参考文件获取详细查询语句。

参考文件	加载时机	核心内容
compute-instances.md	处理Compute Engine虚拟机、机器类型、IP地址、磁盘时	实例清单、机器类型分布、状态检查
kubernetes-gke.md	处理GKE集群、节点池、Pod、部署、服务、RBAC时	集群拓扑、工作负载分布、RBAC绑定
serverless-containers.md	处理Cloud Run服务、修订版本、执行记录时	服务清单、修订版本追踪、执行分析
networking-dns.md	处理VPC网络、子网、路由、DNS记录时	网络拓扑、子网分析、路由表、DNS记录集
messaging-pubsub.md	处理Pub/Sub主题时	主题清单、消息拓扑
iam-security.md	处理IAM服务账号、角色、Secret Manager时	服务账号审计、角色分析、密钥版本追踪
monitoring-logging.md	处理监控仪表板、日志、已保存查询时	仪表板清单、已保存查询分析
resource-management.md	处理GCP项目、区域、组织层级时	项目清单、区域分布、层级映射
resource-ownership.md	处理资源归属、GCP标签、组织结构时	基于标签的分组、项目级汇总、成本分摊

Best Practices

最佳实践

Configuration Parsing

配置解析

Always parse

gcp.object

with JSON parser:

parse gcp.object, "JSON:gcpjson"

Access primary resource attributes via
```
gcpjson[configuration][resource][...]
```

Access extended properties via

gcpjson[configuration][additionalAttributes][...]

Check for null values after parsing with
```
isNotNull()
```

始终使用JSON解析器解析

gcp.object

：

parse gcp.object, "JSON:gcpjson"

通过
```
gcpjson[configuration][resource][...]
```
访问主要资源属性

通过

gcpjson[configuration][additionalAttributes][...]

访问扩展属性

解析后使用
```
isNotNull()
```
检查空值

GCP Hierarchy

GCP层级

Organization → Folder → Project → Region/Zone
Use
```
gcp.project.id
```
as the primary scoping filter
Use
```
gcp.organization.id
```
for cross-project queries
Use
```
gcp.region
```
and
```
gcp.zone
```
for location-based analysis

层级顺序：组织 → 文件夹 → 项目 → 区域/可用区
使用
```
gcp.project.id
```
作为主要范围筛选条件
使用
```
gcp.organization.id
```
进行跨项目查询
使用
```
gcp.region
```
和
```
gcp.zone
```
进行基于位置的分析

Entity Naming

实体命名

Entity types follow the
```
GCP_<SERVICE_API>_<RESOURCE>
```
format
Service API maps to the Google API domain with underscores replacing dots and hyphens
Use specific entity types (avoid
```
"GCP_*"
```
wildcards when possible)

实体类型遵循
```
GCP_<SERVICE_API>_<RESOURCE>
```
格式
服务API对应Google API域名，用下划线替换点和连字符
使用特定实体类型（尽可能避免
```
"GCP_*"
```
通配符）

Labels

Relationship Traversal

关系遍历

Use
```
"*"
```
as the relationship name — GCP entities do not have named relationship types
Use
```
fieldsKeep
```
to maintain important fields through traversal
Access traversal history with
```
dt.traverse.history[-N]
```
Complex topologies may require multiple traverse operations

使用
```
"*"
```
作为关系名称 — GCP实体没有命名的关系类型
使用
```
fieldsKeep
```
在遍历过程中保留重要字段
使用
```
dt.traverse.history[-N]
```
访问遍历历史
复杂拓扑可能需要多次遍历操作

Limitations and Notes

限制与说明

Smartscape Limitations

Smartscape限制

Smartscape data reflects the most recent scan; there may be a delay between GCP changes and Dynatrace visibility
Not all GCP services are represented as entity types
Some configuration fields may be null depending on resource setup
Resource discovery depends on GCP integration configuration

Smartscape数据反映最新扫描结果；GCP变更与Dynatrace可见性之间可能存在延迟
并非所有GCP服务都以实体类型呈现
根据资源设置，部分配置字段可能为空
资源发现依赖GCP集成配置

GCP-Specific Notes

GCP特定说明

GCP labels must be accessed via backtick syntax:
```
`tags:gcp_labels`
```
GCP entities use
```
"*"
```
for relationship traversal (no named relationship types)
GCP object configuration requires parsing with
```
parse gcp.object, "JSON:gcpjson"
```
Configuration fields nest under
```
gcpjson[configuration][resource][...]
```
(differs from AWS pattern)

GCP标签必须使用反引号语法访问：
```
`tags:gcp_labels`
```
GCP实体在关系遍历时使用
```
"*"
```
（无命名关系类型）
GCP对象配置需要使用
```
parse gcp.object, "JSON:gcpjson"
```
解析
配置字段嵌套在
```
gcpjson[configuration][resource][...]
```
下（与AWS模式不同）

General Tips

通用提示

Filter early by project and region for better performance
Use
```
isNotNull()
```
and
```
isNull()
```
for graceful null handling
Combine project and region filters for large environments
Use
```
countDistinct()
```
for unique resource counts
Limit results with
```
| limit N
```
during exploration

尽早按项目和区域筛选以提升性能
使用
```
isNotNull()
```
和
```
isNull()
```
优雅处理空值
在大型环境中组合项目和区域筛选条件
使用
```
countDistinct()
```
统计唯一资源数量
探索时使用
```
| limit N
```
限制结果数量