Failed to aggregate batch

Hello.

Trying to create alert via Chronograf web interface. All queries with 'group by' enabled like

SELECT median("load1") AS "median_load1" FROM "telegraf"."autogen"."system" WHERE time > now() - 15m GROUP BY time(5m)

fails inside kapacitor with

[chronograf-v1-7a28bb6e-f659-40ce-b42b-6ada5711db7c:median3] 2017/04/10 13:44:14 E! failed to aggregate batch: field load1 missing from point cannot aggregate

Direct queries to influx retrun results:

InfluxDB shell version: 1.2.2
> SELECT median("load1") AS "median_load1" FROM "telegraf"."autogen"."system" WHERE time > now() - 15m GROUP BY time(5m)
name: system
time median_load1
---- ------------
1491823800000000000 0.01
1491824100000000000 0
1491824400000000000 0.155
1491824700000000000 0.2

Versions:

- influx: 1.2.2
- kapacitor: 1.2.0
- chronograf: 1.2.0~beta8

Can you give any advice what could be wrong or how I can debug these?

Start by reproducing the grouped median query through Chronograf and Kapacitor using the reported InfluxDB, Kapacitor, and Chronograf versions, then compare it with the successful direct InfluxDB query. Done means identifying why the batch aggregation reports the load1 field as missing and recording or addressing the cause.

Written by the indexing model from the issue text.