{{.ID}} generated with empty values despite of groupBy well formed

Hi:

I'm using kapacitor 1.6.5 with an InfluxDB 1.8.10. The TICK script was generated by Chronograf 1.10.0 (latest from docker).

The problem is that {{.ID}} is not well mapped because empty values. When generating a recording and replaying, what I see is:

`
ts=2022-09-29T09:26:19.919Z lvl=debug msg="alert triggered" service=kapacitor task_master=main task_master=e392cbcd-292c-45cb-b9bf-2405b3acd3cb task=chronograf-v1-40ce5d94-1dad-4a6c-87a8-a5e6b363414e node=alert8 level=CRITICAL id=FAILOVER_detect-device=,kpi=,rg= event_message="[OPENNTI] 2022-09-28 11:37:00 +0000 UTC FAILOVER Fields: map[current.device:SECURGW103-1_MAQUETA current.kpi:chassis.cluster.redundancy-group.redundancy-group-failover-count current.product-model:srx5800 current.rg:2 current.value:5 current.version:20191115.14c2ad5_builder_stable_11 past.device:SECURGW103-1_MAQUETA past.kpi:chassis.cluster.redundancy-group.redundancy-group-failover-count past.product-model:srx5800 past.rg:2 past.value:4 past.version:20191115.14c2ad5_builder_stable_11 value:1] Tags: map[device: kpi: rg:]" data="&{chassis.cluster.redundancy-group.redundancy-group-failover-count map[device: kpi: rg:] [time current.device current.kpi current.product-model current.rg current.value current.version past.device past.kpi past.product-model past.rg past.value past.version value] [[2022-09-28 11:37:00 +0000 UTC SECURGW103-1_MAQUETA chassis.cluster.redundancy-group.redundancy-group-failover-count srx5800 2 5 20191115.14c2ad5_builder_stable_11 SECURGW103-1_MAQUETA chassis.cluster.redundancy-group.redundancy-group-failover-count srx5800 2 4 20191115.14c2ad5_builder_stable_11 1]]}"
`

As you can see, id has the following value:

```
id=FAILOVER_detect-device=,kpi=,rg=
```

But the detection works fine.

{{.ID}} is build in the Chronograf way:

```
var groupBy = ['device', 'kpi', 'rg']
var idVar = name + '-{{.Group}}'

var trigger = past
|join(current)
.as('past', 'current')
|eval(lambda: float("current.value" - "past.value"))
.as('value')
.keep()
|alert()
.crit(lambda: "value" != crit)
.message(message)
.id(idVar)
.idTag(idTag)
.levelTag(levelTag)
.messageField(messageField)
.durationField(durationField)
.noRecoveries()
.stateChangesOnly()
.log('/var/log/kapacitor/FAILOVER_detect.log')
.alerta()
.resource('FAILOVER')
.event('{{.ID}}')
.environment('Production')
.group('tim')
.origin('kapacitor_bentejui')
.services('OPENNTI')
.correlated()
```

Any idea why this behaviour?

The recording was generated with this query:

```
SELECT * FROM "juniper"."four_weeks"."chassis.cluster.redundancy-group.redundancy-group-failover-count" WHERE "device" =~ /SECURGW103/ AND time > now() - 72h
```

If I put a GROUP BY clausule and print all the fields in the query (SELECT "device", "kpi", "rg", last(value) FROM ... GROUP BY "device,"kpi","rg" ... ) the detection fails.

Thanks in advance

Jes

Reproduce the recording and replay using the supplied TICK script and query, then inspect the alert() node's group and tag values as they reach .id('{{.Group}}') and .event('{{.ID}}'). Compare the generated ID with the current, past, and groupBy values shown in the debug output. Done means the behavior is explained and the issue's empty ID values are corrected or documented with a verified reproduction.

Written by the indexing model from the issue text.