Grafana: Difference between revisions
mNo edit summary |
|||
| Line 36: | Line 36: | ||
===Ordering Series=== | ===Ordering Series=== | ||
When making a line graph, I wanted to order series in a specific order. The best way I found was to just have multiple queries selecting specific series in a particular order. | When making a line graph, I wanted to order series in a specific order. The best way I found was to just have multiple queries selecting specific series in a particular order. | ||
=== Tips === | |||
==== Reducing a long chain of conditions ==== | |||
Are you writing Influx queries like: | |||
{{Highlight | |||
| code = SELECT something FROM somewhere | |||
WHERE thing = 'that' OR thing = 'it' OR thing = 'those' OR .... | |||
| lang = sql | |||
}} | |||
Why not group them with pattern matching instead: | |||
{{Highlight | |||
| code = SELECT something FROM somewhere | |||
WHERE thing =~ /^(that{{!}}it{{!}}those{{!}}...)$/ | |||
| lang = sql | |||
}} | |||
Revision as of 00:08, 15 January 2021
Grafana is a tool to generate graphs and dashboards.
Random Notes
Pie Graph using MySQL
With the following query, I am able to generate a two column result set which I want to put into a pie graph.
SELECT requested_cpu, COUNT(requested_cpu) FROM `jobs` GROUP BY requested_cpu ORDER BY `COUNT(requested_cpu)` DESC
However, the Grafana Pie Graph plugin requires the result to be a time series. To work around this, add the timeseries with the current timestamp and ensure that the metric is a char type.
SELECT now() as time_sec, COUNT(requested_cpu) as `value`, CONVERT(requested_cpu, char) as metric
FROM `jobs` WHERE $__timeFilter(started)
GROUP BY requested_cpu
Custom Colors in Pie Graph
Specify custom colors for each series by editing the JSON data directly. There should be an aliasColors field which needs to be populated. For example:
Ordering Series
When making a line graph, I wanted to order series in a specific order. The best way I found was to just have multiple queries selecting specific series in a particular order.
Tips
Reducing a long chain of conditions
Are you writing Influx queries like:
SELECT something FROM somewhere
WHERE thing = 'that' OR thing = 'it' OR thing = 'those' OR ....
Why not group them with pattern matching instead:
SELECT something FROM somewhere
WHERE thing =~ /^(that|it|those|...)$/
