Best practices for writing log files that have variable number of fields

[burwell]

We are writing our own logs for disk usage and we are using key value pairs. The issue is that each host has a different number of disk partitions. So my logs look like the below.

We are not sure what we will do with the data yet. Maybe alert on conditions and maybe collect trending data.
What do people typically do in this case? Thanks.

2017-01-27 02:48:00 db_dt="2017-01-27 02:12:00" hostname=myhost1 vol1 = "/dev/sda1" capacity1 = "706G" percentfull1 = "9%" vol2 = "tmpfs" capacity2 = "7.6G" percentfull2 = "1%" 
2017-01-27 02:48:00 db_dt="2017-01-27 02:12:00" hostname=myhost2 vol1 = "/dev/sda1" capacity1 = "2.4G" percentfull1 = "84%" vol2 = "tmpfs" capacity2 = "24G" percentfull2 = "1%" vol3 = "/dev/sda3" capacity3 = "1.6T" percentfull3 = "1%" 
2017-01-27 02:48:00 db_dt="2017-01-27 02:12:00" hostname=myhost3 vol1 = "/dev/sda1" capacity1 = "474G" percentfull1 = "8%" vol2 = "tmpfs" capacity2 = "12G" percentfull2 = "4%"  vol4=/foo capacity4="3G" percentfull4="17%"

[gcusello]

Hi burwell,
I created for own customer a dashboard (with related alert) that shows all the servers disks, highlighting free space percentage, sending an alert when a disk has a free space percentage less that 10%.
I displayed each disk in a row, reporting in every row also servers informations (hostname, IP, etc...).
In this way I con see all disks, I can filter them and display the ones of only one server, and using colors, I can immediately see the ones in alert.
Putting every disk in a row, I haven't problems for different disks number in my server.
Bye.
Giuseppe