手动创建告警规则需要使用标准 PromQL 语句配置监控对象或指标，本节为您介绍 API 网关支持配置告警的监控指标及每个监控指标对应的 PromQL 语句。

监控指标

实例监控指标

# 平均响应时间
sum(rate(istio_request_duration_milliseconds_sum{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id) / sum(rate(istio_request_duration_milliseconds_count{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id)

# P90 响应时间
histogram_quantile(0.90, sum(rate(istio_request_duration_milliseconds_bucket{apig_io_gateway_id=~".*"}[1m])) by (le, apig_io_gateway_id))

# P99 响应时间
histogram_quantile(0.99, sum(rate(istio_request_duration_milliseconds_bucket{apig_io_gateway_id=~".*"}[1m])) by (le, apig_io_gateway_id))

# HTTP-4XX 错误次数
sum(increase(istio_requests_total{apig_io_gateway_id=~".*",response_code=~"4.*"}[1m])) by (apig_io_gateway_id)

# HTTP-5XX 错误次数
sum(increase(istio_requests_total{apig_io_gateway_id=~".*",response_code=~"5.*"}[1m])) by (apig_io_gateway_id)

# QPS
sum(rate(istio_requests_total{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id)

# HTTP 响应报文大小
sum(rate(istio_response_bytes_sum{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id) / sum(rate(istio_response_bytes_count{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id)

# HTTP 请求报文大小
sum(rate(istio_request_bytes_sum{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id) / sum(rate(istio_request_bytes_count{apig_io_gateway_id=~".*"}[1m])) by (apig_io_gateway_id)

# 入口公网带宽
apig_loadbalancer_in_bytes{apig_io_gateway_id=~".*"}

# 出口公网带宽
apig_loadbalancer_out_bytes{apig_io_gateway_id=~".*"}

# 新建连接数
apig_loadbalancer_new_connections{apig_io_gateway_id=~".*"}

服务监控指标

sum(rate(istio_requests_total{request_service_name=~"app.scer****", apig_io_gateway_id=~"gcek****"}[1m])) by (request_service_name)

# 平均响应时间
sum(rate(istio_request_duration_milliseconds_sum{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name) / sum(rate(istio_request_duration_milliseconds_count{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name)

# P90 响应时间
histogram_quantile(0.90, sum(rate(istio_request_duration_milliseconds_bucket{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (le, request_service_name))

# P99 响应时间
histogram_quantile(0.99, sum(rate(istio_request_duration_milliseconds_bucket{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (le, request_service_name))

# HTTP-4XX 错误次数
sum(increase(istio_requests_total{request_service_name=~".*",response_code=~"4.*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name)

# HTTP-5XX 错误次数
sum(increase(istio_requests_total{request_service_name=~".*",response_code=~"5.*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name)

# QPS
sum(rate(istio_requests_total{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name)

# HTTP 响应报文大小
sum(rate(istio_response_bytes_sum{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name) / sum(rate(istio_response_bytes_count{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name)

# HTTP 请求报文大小
sum(rate(istio_request_bytes_sum{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name) / sum(rate(istio_request_bytes_count{request_service_name=~".*", apig_io_gateway_id=~".*"}[1m])) by (request_service_name)

前提条件

• 已在 VMP 目标工作区下创建告警通知策略，详细操作请参见 创建告警通知策略。
• 已在 VMP 目标工作区下创建告警聚合策略，详细操作请参见 创建告警聚合策略。

操作步骤

1. 登录 API 网关控制台。
2. 在顶部导航栏，选择目标地域。
3. 在左侧导航栏选择 实例管理，单击目标实例名称，进入实例概览页面。
4. 单击 监控信息，切换至 监控信息 页签。
5. 在 监控信息 区域，单击 进入 VMP 管理告警通知，跳转至 VMP 服务控制台的 告警规则 页面。
6. 单击 创建 > 手动创建，为您需要监控的指标配置告警规则。参数介绍可参见 创建告警规则。