Kapacitor alert message not seeing derived fields

kapacitor

#1

I have the below TICK script which is alerting me when network interfaces get saturated over a period of time. (BTW Thanks for the sideload feature)

The last issue that I am having is hopefully a simple one to solve. I am calculating the mean traffic and storing it in the field “net_io” I am able to use that in a comparison later on but I am not able to use this field in my message string.

These lines:

.id('{{ index .Tags "host"}}/{{ index .Tags "interface"}}/net_io')
.message('{{ .ID }}:{{ index .Fields "net_io" }}:{{ index .Fields "net_io_pct" }}')

produce

"id": "spar802.ny1/eth0/net_io",
"message": "spar802.ny1/eth0/net_io:<no value>:0.7986143069514522",

Any idea what I missing here?? Thanks!

// net_io

// DEFINE: kapacitor define net_io -type batch -tick net_io.tick -dbrp telegraf.2years
// ENABLE: kapacitor enable net_io

// Parameters
var warn = 0.70
var crit = 0.75
var period = 5m
var every = 5m

// Dataframe
var data = batch
  |query('''
         SELECT non_negative_derivative(max("bytes_sent"), 1s) + non_negative_derivative(max("bytes_recv"), 1s) AS net_io
         FROM "telegraf"."2years"."net"
         WHERE time < now()
         GROUP BY time(15s)
         ''')
    .period(period)
    .every(every)
    .groupBy(time(15s),'host','interface')
  |mean('net_io')
    .as('net_io')
  |groupBy('host','interface')

var sideload_data = data
  |sideload()
    .source('file:///var/lib/kapacitor/sideload/net_io')
    .order('{{.host}}/{{.interface}}.json')
    .field('speed', 1000)

// Thresholds
var alert = sideload_data
  |eval(lambda: (("net_io" / 1048576.0) * 8.0) / float("speed") )
    .as('net_io_pct')
  |alert()
    .id('{{ index .Tags "host"}}/{{ index .Tags "interface"}}/net_io')
    .message('{{ .ID }}:{{ index .Fields "net_io" }}:{{ index .Fields "net_io_pct" }}')
    .warn(lambda: "net_io_pct" > warn)
    .crit(lambda: "net_io_pct" > crit)

// Alert
alert
  .log('/tmp/net_io_alerts.json')
//  .idField('id')
  .levelTag('level')
  |where(lambda: "level" != '')
  |influxDBOut()
    .database('nagios')
    .retentionPolicy('2years')
    .measurement('net_io')

#2

What i tryied in similar situations and worked for me is to add the .keep(‘net_io’) function to have the values displayed on the messages. Give it a try!

|eval(lambda: ((“net_io” / 1048576.0) * 8.0) / float(“speed”) )
.as(‘net_io_pct’)
.keep(‘net_io_pct’, ‘net_io’)

Hope it works for you as well!


#3

Thanks very much - That worked as it should.


#4

You’re welcome!!

I had this kind of issue with the keep one and also the groupBy…No tags avilable when you forget them on the groupBy side…Same with the value itself and the keep…

But, as i was not sure, i just give you with my little experience on this trial/error matter!!!

Nice to hear that now is working!!