Frequently Asked Questions¶

Why does the scheduled report email return a 414 error?

The image links in the final scheduled report sent to the target email are generated by concatenating the corresponding chart iteration view variables and other parameters of the dashboard. If the content of the view variables is too long, it may cause the generated image link to be too long, resulting in a 414 error.

Why can't the JSON format be used in the Explorer search bar?

Using JSON search requires the following three conditions:

1. The site must be "China Region 1 (Hangzhou)", "China Region 3 (Zhangjiakou)", or "China Region 4 (Guangzhou)"
2. The workspace must be created after June 23, 2022
3. It can only be used in the Log Explorer

How to convert the Grafana dashboard json template provided by Guance to the Guance dashboard JSON template using a node script?

npm download:

npm install  @cloudcare/guance-front-tools

Usage:

$ npm install -g @cloudcare/guance-front-tools

# show usage information
$ grafanaCovertToGuance

# run task
$ grafanaCovertToGuance -d examples/grafana.json -o examples/guance.json

Keywords:

none

Explanation of the P99 peak calculation logic

Guance obtains all values of a certain metric over a period of time, sorts these values from smallest to largest, and then determines the position of the P99 data point based on the number of data points * 99%, returning the value at that position. Here, "all values" refer to the original data points stored in the database.

For percentile queries of metric data, the processing method of Guance varies depending on the data engine used. If the database itself does not support percentile queries, Guance will pull the data locally and perform additional processing to achieve this function. For non-metric data (such as logs, APM, RUM, etc.), percentile queries are supported by the native functions of the database.