regel

{
  "type": "timeseries",
  "name": "orders-weather",
  "default_datasource": "influx",
  "max_evals": 50,
  "bucket_interval": "1h",
  "span": "auto",
  "offset": 30,
  "interval": "24h",

regel

{
  "type": "timeseries",
  "name": "orders",
  "default_datasource": "influx",
  "max_evals": 50,
  "bucket_interval": "24h",
  "span": "auto",
  "offset": 30,
  "interval": 60,

regel

select temp, clouds from weather where time >='2015-06-30T00:00:00Z' limit 5
name: weather
time                temp               clouds
----                ----               ------
1435622400000000000 18.33000000000004  0
1435626000000000000 17.760000000000048 0
1435629600000000000 16.439999999999998 0
1435633200000000000 15.840000000000032 0
1435636800000000000 15.340000000000032 0

regel

select * from orders where time >='2015-06-30T00:00:00Z' limit 5
name: orders
time                count
----                -----
1435622400000000000 388
1435626000000000000 384
1435629600000000000 375
1435633200000000000 371
1435636800000000000 368

regel

var model = 'telegraf_metrics_count_value_10s'

var from_measurement = 'loudml'

var out_db = 'telegraf'

var retention_policy = 'autogen'

var out_measurement = 'predicted'

regel

var pos = data
    |changeDetect('is_anomaly')
    |alert()
        .warn(lambda: "is_anomaly" == TRUE)
        .crit(lambda: "is_anomaly" == TRUE AND "score" > 90.0)
        .message('{{if ne .Level "OK" }} Hey, unexpected situation detected score={{ index .Fields "score" | printf "%0.3f" }} {{else}} Back to normal {{end}}')
        .pagerDuty2()

regel

data_dir = "/var/lib/kapacitor"

[pagerduty2]
  enabled = true
  routing-key = ""
  url = "https://events.pagerduty.com/v2/enqueue"
  global = false
  
[logging]
    # Destination for logs

regel

var neg = data
    |stateCount(lambda: "is_anomaly" == FALSE)
    |alert()
        .info(lambda: "state_count" == 1)
        .message('{{ .Time }}: situation back to normal model={{ index .Fields "model" }} state=finished score={{ index .Fields "score" | printf "%0.3f" }}.')
        .slack()

regel

var pos = data
    // the next node counts abnormal data points
    |stateCount(lambda: "is_anomaly" == TRUE)
    |alert()
    // Warn after 1 point
        .warn(lambda: "state_count" >= 1)
    // Critical after 5 points and abnormal higher than 90 / 100
        .crit(lambda: "state_count" >= 5 AND "score" > 90.0)
        .message('{{ .Time }}: Hey, unexpected situation detected by model={{ index .Fields "model" }} state=ongoing score={{ index .Fields "score" | printf "%0.3f" }}.')
        .slack()

regel

var data = stream
    // this from node selects all input data received in the given measurement
    |from()
        .measurement(from_measurement)
        .where(lambda: "model" == model)
    // this eval node remembers the model name
    |eval(lambda: model)
        .as('model')
        .keep()
    // the next five lines save the data to the target database into a new measurement