Grok can receive bulk metric data from an uploaded worksheet (.csv) file. The file must be formatted correctly. Once received, data can be imported into Grok as a custom metric and also analyzed for anomalies.
- Click "Upload .csv file" and choose a file. Most file format errors should result in an error. If successful a table will appear with a sample set of rows from your worksheet.
- Verify from the sample that your data is correct. Click "Refresh Rows" to preview different rows.
- Click "Create Metric"
- In the popup, Metric Name must be entered, the other two fields, Group Name and Unit Value are optional. Metric Name must be a url-safe string of text.
- Click "Just Create" to just import the metric into Grok for now. Click "Create and Analyze" to import the metric as well as begin analyzing for anomalies.
- If successful, see results on the Manage Sources page.
Worksheet files should be .csv files with each data row in the format:
Worksheets with empty values, empty rows, delimiters other than commas, and other oddities will not be uploaded correctly. Other restrictions are discussed below.
The datetime field must be in a supported format. Grok will try to detect an ISO 8601 datetime format and also supports UNIX timestamps.
Example ISO 8601 format:
Example Unix Timestamp (seconds since January 1, 1970 UTC) of above date:
Datetime fields must be in sequential order. If a datetime is "before" a previously received datetime, then the import will not be allowed.
The value can be any number, however it must be a number. Any strings or characters other than numbers will cause an error and will not be received or analyzed.
- File too large: At this time, we recommend uploading at most .csv files with a maximum of 50000 entries to ensure that you'll get results in a timely, efficient manner. We will work to improve this as time goes on.
- Unit boundaries: You must set either a min AND max or no values. Most use cases do not require a min/max value.
- Create & Analyze: Sometimes this will not work and throw on error. We recommend just creating the metric first, then visiting the Custom Sources page to analyze the data (and provide the meta data for viewing on the dashboard.