Topbeat Dashboard shows strange numbers in RC1

[owulff]

Hi all

I've upgraded to RC1. I updated the template, deleted old indexes and re-created the index pattern in Kibana. Finally I updated the Beats dashboard.

I liked to see % for filesystem usage and GB, MB etc in Beta 4 but it's not there anymore in RC1.

Further, I think the visualized data doesn't show the reality. I've attached a screenshot and a one topbeat message which shows that the memory usage is close to 100% (It's only one message but it's not a peak it constantly consumes almost 100% memory

{"_index": "topbeat-2015.11.06",
"_type": "system",
"_id": "AVDcVZJK2oFebqHhuu9O",
"_score": 1.6216148,
"_source": {
"@timestamp": "2015-11-06T10:26:45.972Z",
"count": 1,
"cpu":
{"user": 1117722847,
"user_p": 0.05,
"nice": 18000892,
"system": 29156815,
"system_p": 0.01,
"idle": 1772013661,
"iowait": 9352671,
"irq": 895289,
"softirq": 11469696,
"steal": 0},
"load":
{"load1": 0.07,
"load5": 0.2,
"load15": 0.26},
"mem":
{"total": 8254984192,
"used": 8114094080,
"free": 140890112,
"used_p": 0.98,
"actual_used": 7181971456,
"actual_free": 1073012736,
"actual_used_p": 0.87},
"shipper": "DEVLX5483",
"swap":
{"total": 2147475456,
"used": 525570048,
"free": 1621905408,
"used_p": 0.24,
"actual_used": 0,
"actual_free": 0,
"actual_used_p": 0},
"tags":
["beats"],
"type": "system",
"@version": "1"}}

What am I missing?

Thanks
Oli

[tudor (Tudor Golubenco)]

What ES and Kibana versions are you using?

The %, MB, GB, etc. are done via field formatters in Kibana, and they should be configured automatically when you load the beats-dashboards. Can you check the index/fields settings in Kibana? This is how it looks for me:

[owulff]

My mistake. I didn't notice that you upload the index pattern as well in the script. I've re-created that without the format

Nevertheless, I still see strange numbers in the dashboards.

Memory usage is close to 100% (which is correct based on the uutput of top). Nevertheless, the visualization "Memory per process" it goes up to 25% only. Also "top ten memory consumers" doesn't make too much sense if memory usage is at almost 100%.