Business-Intelligent Observability Service (BOS) provides the custom monitoring feature that allows you to use monitoring templates to visualize custom metrics.
The custom monitoring feature allows you to flexibly split log content and aggregate data by dimensions. You can use this feature to perform real-time business analysis and configure alert rules based on the business logs of applications. This way, data can be monitored in a way as easy as by using SQL statements. You can use this feature to monitor personalized and non-standard business metrics, such as transaction volumes.
The following figure shows the process of custom processing, which consists of data source selection, data filtering, statistic collection, and alert configuration.
For example, you can use the custom monitoring feature to obtain the number of transactions that are created for each product in each minute based on the following sample logs:
2012-11-11 11:11:11,295 0211004, Clothes, Deal, Create, 7,Y,
2012-11-11 11:11:12,295 0211005, Clothes, Transaction, Creation, 8,Y,
2012-11-11 11:11:13,295 0211006, Food, Transaction, Creation, 9,Y,
2012-11-11 11:11:14,295 0211007, Home, Transaction, Creation, 11,Y,
2012-11-11 11:11:15,295 0211008, Food, Trade, Creation, 7,N,
……Notes:
,is the delimiter that is used to split fields in the sample logs. Each log is displayed in the following format:Log creation time,Order ID,Product, Service name,Method name,Time consumed,Result.Statistic collection: RMS monitors massive data. Statistics on raw structured data need to be collected to facilitate data monitoring. Common statistical methods include calculating the number of rows, the value sum, the average value, the largest value, and the smallest value.
Alert configuration: You can define various alert rules based on statistics such as an absolute value, a same-period-last-cycle comparison, a period-over-period comparison, and a sum for the last N minutes. RMS allows you to define multiple rules and logical operations at a time.