CloudMonitor, Log Service, and Cost Management

Understanding how to monitor your Alibaba Cloud resources and manage costs is just as important as knowing how to build with them. This lesson covers CloudMonitor, Log Service, ActionTrail, and billing tools.

---

CloudMonitor

CloudMonitor is Alibaba Cloud's primary monitoring service. It collects metrics from your cloud resources and enables alerting.

Metrics

CloudMonitor automatically collects metrics from Alibaba Cloud services:

Service	Example Metrics
ECS	CPU utilisation, memory usage, disk I/O, network traffic
RDS	Connections, IOPS, storage usage, replication delay
OSS	Request count, bandwidth, error rate
SLB	Active connections, request rate, latency
Function Compute	Invocations, duration, errors

Alert Rules

Define alert rules to notify you when metrics exceed thresholds:

• Threshold alerts — trigger when a metric exceeds a value for a specified duration
• Contact groups — define who receives alerts (email, SMS, DingTalk)
• Escalation — automatically escalate if alerts are not acknowledged

Example alert: If ECS CPU utilisation exceeds 80% for 5 consecutive minutes, send an alert to the Operations team via DingTalk.

Dashboards

Create custom dashboards to visualise metrics:

• Combine metrics from multiple services
• Use line charts, bar charts, gauges, and tables
• Share dashboards with team members

Host Monitoring

Install the CloudMonitor agent on ECS instances for detailed OS-level metrics:

• Memory utilisation (not available via basic monitoring)
• Disk utilisation and I/O wait
• TCP connection count
• Process-level monitoring

---

Log Service (SLS)

Log Service (SLS) is a fully managed platform for log collection, processing, and analysis. It is the equivalent of AWS CloudWatch Logs + OpenSearch combined.

Key Capabilities

Capability	Description
Log Collection	Collect logs via Logtail agent, SDKs, Kafka protocol, or Syslog
Log Query	SQL-like query language with sub-second response
Log Analysis	Built-in charts, dashboards, and anomaly detection
Alerting	Trigger alerts based on log query results
Data Shipping	Forward logs to MaxCompute, OSS, or AnalyticDB

Log Projects and Logstores

• Project — a container for logstores, dashboards, and alert rules
• Logstore — a unit of log storage and querying (similar to a log group)

Query Syntax

SLS supports both keyword search and SQL analytics:

-- Keyword search
error AND status: 500

-- SQL analytics
* | SELECT status, COUNT(*) AS cnt
     FROM log
     WHERE __time__ > UNIX_TIMESTAMP() - 3600
     GROUP BY status
     ORDER BY cnt DESC

Common Log Sources

Source	Collection Method
ECS instances	Logtail agent
Kubernetes containers	Logtail DaemonSet
Function Compute	Automatic
API Gateway	Built-in integration
OSS access logs	Enable access logging
VPC Flow Logs	Built-in integration

---

ActionTrail