Metrics Monitoring and Alerting System

These flashcards cover key concepts and parameters from the metrics monitoring and alerting system lecture notes.

What is the primary goal of a metrics monitoring and alerting system?

To ensure high availability and reliability.

What types of metrics should be collected in the designed system?

Operational metrics such as CPU load, Memory, and Data disk space.

What is the scale of the infrastructure discussed in the notes?

100 million daily active users, 1000 server pools, and 100 machines per pool.

How long should data be retained in the system?

Data retention should be for 1 year.

What is the data resolution strategy for long-term storage?

Keep newly received metrics for 7 days, roll them up to 1m resolution for the next 30 days, and further roll them up to 1h resolution after 30 days.

What are the supported alert channels mentioned in the chapter?

Email, phone, PagerDuty, or webhooks.

What are the five core components involved in a metrics monitoring and alerting system?

Data collection, Data transmission, Data storage, Alerting, Visualization.

What type of data model is used for metrics data recording?

Time-series data model, which includes a set of values with timestamps.

What is the recommendation regarding the type of database for storing metrics data?

It is not recommended to use a general-purpose database; dedicated time-series databases are preferable.

Name two popular time-series databases mentioned in the notes.

InfluxDB and Prometheus.