Management Console:Appendix: Products supported by health status

Category

Product name

Abbreviation

Resource type

Definition of unavailability

Unavailability criteria

Interval

Compute

ECS

ECS

instance

An ECS instance is unavailable during a specific period if its performance is affected or degraded due to limitations of the underlying host, storage, or network. This includes issues such as Out of Memory (OOM) errors, blue screens, or hangs. When this happens, the system marks its status as abnormal. You can restart the instance to restore service, or submit a ticket to report the failure.

ECS monitors the instance status and reports events to CloudMonitor.

Product: ecs

Event name: Instance:HealthStatusChange. Unavailability is indicated if healthStatus changes to Impaired.

Recorded on each status change

Function Compute

FC

service

An FC service is unavailable if its request error rate (requests with an HTTP 500 status code / total requests) exceeds 10% during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

FC monitors service requests and reports time-series metrics to CloudMonitor.

Namespace: acs_fc

Failed requests: ServiceServerErrors

Error rate: ServiceServerErrorsRate

Unavailability is indicated if ServiceServerErrorsRate > 10%. However, if ServiceServerErrors is less than 10 per minute, the error rate is not calculated, and the service is considered normal by default.

Recorded every 10 minutes (fixed interval)

Storage

Object Storage Service

OSS

bucket

An OSS bucket is unavailable if the HTTP request error rate (failed requests with a 5xx status code / total requests) exceeds 10% during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

OSS monitors and calculates HTTP requests and reports time-series metrics to CloudMonitor. Namespace: acs_oss_dashboard

Availability metric: Availability

Unavailability is indicated if Availability < 90%.

Recorded every 10 minutes (fixed interval)

Elastic Block Storage

EBS

disk

An EBS disk is unavailable if it cannot be read from or written to, or if its read/write performance is severely degraded during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

EBS monitors the disk status and reports system events to CloudMonitor.

Product: ecs

Local disk corruption alert starts: Disk:ErrorDetected:Executing

Disk performance severely affected: Disk:Stalled:Executing

Unavailability is indicated if either of these two events occurs.

Recorded on each status change

Log Service

SLS

log project

A request that fails due to an SLS issue is a failed request. The service is unavailable if the error rate (failed requests / valid requests) exceeds 10% for 10 consecutive minutes. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket.

SLS monitors service requests and reports time-series metrics to CloudMonitor.

Namespace: acs_sls_dashboard

Failed requests: The number of 5xx errors filtered from the LogCodeQps metric.

Valid requests: The total count of LogCodeQps.

Unavailability is indicated if the error rate (total failed requests / total valid requests) in a 10-minute period > 10%. If the total number of valid requests is < 100, the error rate is not calculated, and the service is considered normal by default.

Recorded every 10 minutes (fixed interval)

Networking & CDN

Content Delivery Network

CDN

domain name

A CDN service is unavailable if its request error rate (failed requests with a 5xx status code returned by the domain name due to CDN system issues / total requests) exceeds 10% during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

CDN monitors service requests and reports time-series metrics to CloudMonitor. Namespace: acs_cdn

Error rate metric: code5xx

Total requests: QPS*60. Unavailability is indicated if code5xx > 10%. However, if the total number of requests per minute (QPS*60) is less than 40, the error rate is not calculated, and the service is considered normal by default.

Recorded every 10 minutes (fixed interval)

Dynamic Content Delivery Network

DCDN

domain name

A DCDN service is unavailable if its request error rate (failed requests with a 5xx status code returned by the domain name due to DCDN system issues / total requests) exceeds 10% during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

DCDN monitors service requests and reports time-series metrics to CloudMonitor. Namespace: acs_dcdn

Error rate metric: code_ratio_5

Total requests: dcdn_qps*60. Unavailability is indicated if code_ratio_5 > 10%. However, if the total number of requests per minute (dcdn_qps*60) is less than 40, the error rate is not calculated, and the service is considered normal by default.

Recorded every 10 minutes (fixed interval)

NAT Gateway

NAT Gateway

instance

A NAT Gateway instance is unavailable if all its outbound data packets are dropped by the Alibaba Cloud egress gateway device during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket.

NAT Gateway monitors the instance status and reports system events to CloudMonitor.

Product: nis

NAT data path unavailable event: problem-nat-datapathUnavailable

An event is sent if the unavailability lasts for more than 10 minutes. The event is re-sent every 10 minutes if the issue persists. The service is considered recovered when these events stop.

Recorded every 10 minutes (arbitrary interval)

Middleware

API Gateway

API Gateway

instance

An API Gateway instance is unavailable if its request error rate (failed API calls due to gateway issues / total requests) exceeds 10% during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

API Gateway monitors and calculates service requests and reports time-series metrics to CloudMonitor. Namespace: acs_apigateway_dashboard

Total requests: request_count_instance

Failed requests: error_count_instance

Instance node loss ratio: instance_lost

1. The instance is unavailable if instance_lost = 100%.

2. The instance is unavailable if (error_count_instance / request_count_instance) > 10%. However, if request_count_instance is less than 100 per 10 minutes, the error rate is not calculated, and the service is considered normal by default.

Recorded every 10 minutes (fixed interval)

ApsaraMQ for RocketMQ

ApsaraMQ

instance

An ApsaraMQ for RocketMQ instance is unavailable if all attempts to connect to it fail, or if all API requests to produce, consume, or confirm messages fail during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket. (Note: Only RocketMQ 5.x instances are supported.)

ApsaraMQ for RocketMQ monitors instance status and reports availability changes through CloudMonitor system events.

Product: rocketmq

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

ApsaraMQ for Kafka

Kafka

instance

An ApsaraMQ for Kafka instance is unavailable if all continuous attempts to call the Topic API of the specified instance fail, or if message delivery latency exceeds 5 minutes. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket.

ApsaraMQ for Kafka monitors instance status and reports availability changes through CloudMonitor system events.

Product: kafka

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

Database

ApsaraDB for RDS

RDS

instance

An RDS instance is unavailable if all continuous attempts to connect to it fail during a period. When this happens, the system marks its status as abnormal. You can try to restart the instance to restore service or submit a ticket to report the failure.

RDS monitors the instance status and reports system events to CloudMonitor.

Product: rds;

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

ApsaraDB for MongoDB

MongoDB

instance

An ApsaraDB for MongoDB instance is unavailable if all continuous attempts to connect to it fail during a period. When this happens, the system marks its status as abnormal. You can try to restart the instance, component, or node to restore service or submit a ticket to report the failure.

ApsaraDB for MongoDB monitors the instance status and reports system events to CloudMonitor.

Product: mongodb;

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

PolarDB

PolarDB

cluster

A PolarDB cluster is unavailable if all continuous attempts to connect to it fail during a period. When this happens, the system marks its status as abnormal. You can try to use node management to restore service or submit a ticket to report the failure.

PolarDB monitors the instance status and reports system events to CloudMonitor.

Product: polardb;

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

ApsaraDB for Redis

Redis

instance

An ApsaraDB for Redis instance is unavailable if all continuous attempts to connect to it fail during a period. When this happens, the system marks its status as abnormal. You can try to restart the instance to restore service or submit a ticket to report the failure.

ApsaraDB for Redis monitors the instance status and reports system events to CloudMonitor.

Product: redis;

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

PolarDB-X

PolarDB-X

instance

A PolarDB-X instance is unavailable if all continuous attempts to connect to it fail during a period. When this happens, the system marks its status as abnormal. You can try to restart the instance to restore service or submit a ticket.

PolarDB-X monitors the instance status and reports system events to CloudMonitor.

Product: drds

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

AnalyticDB for PostgreSQL

AnalyticDB for PostgreSQL

instance

An AnalyticDB for PostgreSQL instance is unavailable if all continuous attempts to connect to it fail during a period. When this happens, the system marks its status as abnormal. You can try to restart the instance to restore service or submit a ticket.

AnalyticDB for PostgreSQL monitors instance status and reports availability changes through CloudMonitor system events.

Product: gpdb

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

Data Transmission Service

DTS

instance

A DTS instance is unavailable if data written to it cannot be synchronized to the destination and this status persists for more than 5 minutes. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket.

DTS monitors instance status and reports time-series metrics to CloudMonitor.

Namespace: acs_dts

Synchronization task instance unavailable: Unavailability is indicated if SynchronizationLatency (synchronization task latency) exceeds 5 minutes and persists for more than 5 minutes.

Subscription task instance unavailable: Unavailability is indicated if ChangeTrackingLatency (subscription task latency) exceeds 5 minutes and persists for more than 5 minutes.

Migration task instance unavailable: Unavailability is indicated if MigrationLatency (migration task latency) exceeds 5 minutes and persists for more than 5 minutes.

Recorded every 5 minutes (fixed interval)

Lindorm

Lindorm

instance

A Lindorm instance is unavailable for a given minute if all attempts to connect to it fail within that minute. When this happens, the system marks its status as abnormal.

Lindorm monitors instance status and reports system events to CloudMonitor.

Product: hitsdb;

Event names: Instance_Failure_Start (failure starts, unavailable); Instance_Failure_End (failure ends, normal).

Recorded on each status change

AI & Machine Learning

OpenSearch

OpenSearch

application

An OpenSearch application is unavailable if all continuous attempts to initiate search requests to it fail during a period. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket.

OpenSearch monitors and calculates service requests and reports time-series metrics to CloudMonitor.

Namespace: acs_opensearch

Request status: StatusCodeByGateway, where 5xx indicates a failed request.

Unavailability is indicated if all requests within a 10-minute period return a 5xx status code and the total number of requests is > 30.

Recorded every 10 minutes (fixed interval)

Others

ApsaraVideo Live

LIVE

domain name

A request is considered failed due to an ApsaraVideo Live fault if: (1) for FLV-HTTP/RTMP, no data is returned for 20 seconds while the connection is maintained; or (2) for HLS, a 5xx error occurs. The service is unavailable if the error rate (failed requests / total requests) exceeds 20%. When this happens, the system marks its status as abnormal. If the service does not recover promptly, you can submit a ticket to report the failure.

ApsaraVideo Live monitors service requests and reports time-series metrics to CloudMonitor. Namespace: acs_videolive

Error rate metric: code5xx

Total requests: QPS*60

Unavailability is indicated if code5xx > 20%. However, if the total number of requests per minute (QPS*60) is less than 100, the error rate is not calculated, and the service is considered normal by default.

Recorded every 10 minutes (fixed interval)