Updated on 2024-08-05 GMT+08:00

Resource Monitoring

Table 1 Basic concepts

Terminology

Description

Metrics

Metrics reflect resource performance data or status. A metric consists of a namespace, dimension, name, and unit.

Metric namespaces can be regarded as containers for storing metrics. Metrics in different namespaces are independent of each other so that metrics of different applications will not be aggregated to the same statistics information. Each metric has certain features, and a dimension may be considered as a category of such features.

Host

Each host of AOM corresponds to a VM or physical machine. A host can be your own VM or physical machine, or an Elastic Cloud Service (ECS) or Bare Metal Server (BMS) purchased. A host can be connected to AOM for monitoring only when its OS meets requirements and it is installed with an ICAgent.

Logs

You can quickly search for required logs from massive quantities of logs. You can also quickly locate faults by analyzing the log source and context.

Log traffic

Log traffic refers to the volume of logs reported per second. A maximum of 10 MB/s is supported for each tenant in a region. If the log traffic exceeds 10 MB/s, logs may be lost.

Alarms

Alarms are reported when AOM, ServiceStage, APM, or CCE is abnormal or may cause exceptions. Alarms will cause service exceptions and need to be handled.

Events

Events generally carry important information. They are reported when AOM, ServiceStage, APM, or CCE encounters some changes. Events do not necessarily cause service exceptions. Events do not need to be handled.

Alarm clearance

There are two alarm clearance modes:

  • Automatic clearance: After a fault is rectified, AOM automatically clears the corresponding alarm.
  • Manual clearance: After a fault is rectified, AOM does not automatically clear the corresponding alarm. Instead, you need to manually clear the alarm.

Alarm rules

Alarm rules are classified into metric alarm rules and event alarm rules.

  • Metric alarm rules monitor the usage of resources (such as hosts and components) in the environment in real time.
  • If there are many resource alarms but you do not want to receive notifications too often, set event alarm rules to quickly identify specific types of resource usage problems.

Alarm notification

There are two alarm notification modes:

  • Direct alarm reporting: When setting alarm notification rules, specify alarm notification recipients so that they can take measures to rectify faults in a timely manner. Alarms can be sent through email and SMS.
  • Alarm noise reduction: Select a grouping rule to reduce alarm noise.

Alarm action rules

An alarm action rule defines the action to be taken after an alarm is generated. It includes where the message is sent and in what form.

Prometheus instances

Logical units used to manage Prometheus data collection, storage, and analysis.

Prometheus probes

Deployed in the Kubernetes clusters on the user or cloud product side. Prometheus probes automatically discover targets, collect metrics, and remotely write data to databases.

Exporters

Collect monitoring data and regulate the data provided for external systems using the Prometheus monitoring function. Currently, hundreds of official or third-party exporters are available. For details, see Exporters.

Jobs

Configuration set for a group of targets. Jobs specify the capture interval, access limit, and other behavior for a group of targets.