Here, we will use Timelion to apply conditional logic on metricdata time-series data to track the trends with a moving average. Using this process, we can easily detect the patterns and outliers in a time series.
For this practical exercise, I will be using the metricbeat data, and to explain this process, we will pick the system.memory.actual.used.bytes field of data. So, first of all, I will plot the system memory used in bytes with the following command:
.es(index=metricbeat-*, timefield='@timestamp', metric='max:system.memory.actual.used.bytes').label('Memory used in bytes')
In the preceding command, I am using the .es() function, which is used to query with Elasticsearch index. The first parameter of the query ...