ZStack Cloud 4.8.10>ZStack Cloud Product Manuals>CLI Command Manual>Platform O&M>Cloud Monitoring

Platform O&M

Cloud Monitoring

Overview

ZWatch

ZWatch mainly provides the following features:

Time-series monitoring. Currently, ZWatch monitors the following two types of time-series data (metrics):
- Resource load, such as the VM instance CPU utilization and host memory utilization.
- Resource capacity, such as the available IP count and total count of running VM instances. Metrics (time series data) refers to data collected at a specified interval. Each metric is associated with a timestamp field.
Event collection. ZWatch collects predefined events in the system, such as the host disconnection and VM instance HA enabling.
Alarm reporting. Currently, ZWatch supports the following two types of alarms:
- Metric alarm. For example, you can set an alarm for VM instance CPU utilization, trigger the alarm when the CPU utilization exceeds 80% for five consecutive seconds, and send alarm messages to an email address.
- Event alarm, also known as event subscription. For example, you can subscribe to the host disconnection event and send a notification to the administrator via an email if a host is disconnected.
Auditing. ZWatch records all API operations for your subsequent search and query.
Metric customization. ZWatch allows you to call ZWatch APIs to write metrics. The metrics you wrote can be used for query and alarm setting.
Text template (alarm message template) customization. ZWatch allows you to customize alarm message templates as needed.

SNS

One major feature of the Simple Notification Service (SNS) is to push messages to the endpoints that you subscribed to. The endpoints include email, HTTP webhook, DingTalk, and Short Message Service (SMS). ZWatch directly uses SNS to push alarm messages to users.

Basic Workflow

Create an SNS topic to receive alarm messages.
Create an email platform if you use an email endpoint. Skip this step if you use the HTTP webhook, DingTalk, or SMS endpoint.
Create an endpoint (email, HTTP webhook, DingTalk, or SMS) and subscribe the endpoint to the SNS topic you created in Step 1.
Note: Related SNS resources mentioned above can be reused by multiple alarms.
Create an alarm and set Action to the SNS topic you created in Step 1.

ZWatch

The following terminology and concepts are central to your understanding and use of ZWatch:

Metric: a set of time-series data published to ZWatch, such as the VM CPU utilization.
Event: a set of discrete data published to ZWatch, such as host disconnection.
Note:
- A metric is a set of time-series data, such as the VM CPU utilization collected at an interval of 10 seconds.
- An event is a set of discrete data, such as the host disconnection event occurred after network disconnection.
Alarm: a metric alarm that can be triggered under specific condition. For example, if the CPU utilization of a particular VM instance exceeds 80% in 5 consecutive minutes, an alarm can be triggered and alarm messages can be sent to users via email.
Event Subscription: an alarm that is only applicable to events. For example, if the host disconnection event occurs, an event subscription can be sent to users via email.
Text Template: a template for formatting alarm texts when an alarm or event subscription is triggered.

Next Chapter: Backup Management

Platform O&M

Cloud Monitoring

Overview

ZWatch

SNS

Basic Workflow

ZWatch

Products

Solutions

Help & Support

Contact Us