Help Center/
Well-Architected Framework/
Well-Architected Framework and Practices/
Operational Excellence Pillar/
OPS07 Fault Analysis and Management/
OPS07-01 Creating Alarms
Updated on 2025-05-22 GMT+08:00
OPS07-01 Creating Alarms
- Risk level
High
- Key strategies
Respond to alarms based on the alarm operability principle. You do not need to respond to meaningless alarms, like a sudden increase in disk I/O volume or CPU usage. Following the operability principle can prevent many false alarms. In addition, you need to periodically collect statistics on and analyze alarm frequencies, identify high-frequency alarms, resolve the alarms, and clear the false alarms.
- Design suggestions
- Optimize alarm thresholds: Properly increase the memory, CPU, and network I/O alarm thresholds.
- Optimize log levels: Optimize improper log levels, for example, change some ERROR logs to WARNING logs.
- Shield some logs: For applications whose log levels are difficult to adjust, shield some frequent log alarms based on keywords.
- Enhance warning: Provide warnings for some operations that affect service parties.
- Enhance emergency warning: Some hardware faults will occur in /var/log/messages. Match hardware alarms based on keywords for timely handling.
- Related cloud services and tools
Parent topic: OPS07 Fault Analysis and Management
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
The system is busy. Please try again later.
For any further questions, feel free to contact us through the chatbot.
Chatbot
