Advanced Watch of serial diff of value

[6tUBp3gwd9Zp]

I want to create an advanced watcher that monitors a value and if that value does make a jump greater than x then i want to be informed. My guess was to do this with a serial_diff of the average aggregation of that value and compare that to a treshold but i cant get it working. Is it even possible? And if so, how?

[Bill_McConaghy]

Moved to Elasticsearch topic as this is really an ES question.

[6tUBp3gwd9Zp]

So I solved it like this. Maybe it's helpful for others. Nevertheless other maybe better solutions are much appreciated.

{
"metadata": {
	"..."
},
"trigger": {
	"..."
},
"input": {
	"search": {
		"request": {
			"search_type": "query_then_fetch",
			"indices": ["...*"],
			"types": ["doc"],
			"body": {
				"size": 0,
				"aggs": {
					"1": {
						"max_bucket": {
							"buckets_path": "1-bucket>1-metric"
						}
					},
					"1-bucket": {
						"date_histogram": {
							"field": "robotime",
							"interval": "1h",
							"time_zone": "Europe/Berlin",
							"min_doc_count": 0
						},
						"aggs": {
							"avg_offset": {
								"avg": {
									"field": "property.gapMeasurementResultInMm.leftOffset"
								}
							},
							"1-metric": {
								"serial_diff": {
									"buckets_path": "avg_offset"
								}
							}
						}
					}
				},
				"query": {
					"..."
				},
			}
		}
	}
},
"condition": {
	"script": {
		"source": "if (ctx.payload.aggregations.1.value > params.threshold) { return true; } return false;",
		"lang": "painless",
		"params": {
			"threshold": 0.5
		}
	}
},
"actions": {
	"..."
}
}