Aggregating data with multiple fields

I'm trying to aggregate all fields of a certain series to time steps of 1h. With continuous queries, this worked as follows:

```
CREATE CONTINUOUS QUERY cq_1h_net_sec ON telegraf
BEGIN
SELECT
mean(bytes_sec_recv) as bytes_sec_recv_avg,
max(bytes_sec_recv) as bytes_sec_recv_max,
min(bytes_sec_recv) as bytes_sec_recv_min,
percentile(bytes_sec_recv, 95) as bytes_sec_recv_perc95,
mean(bytes_sec_sent) as bytes_sec_sent_avg,
max(bytes_sec_sent) as bytes_sec_sent_max,
min(bytes_sec_sent) as bytes_sec_sent_min
percentile(bytes_sec_sent, 95) as bytes_sec_sent_perc95,
INTO "store1y"."net_sec"
FROM "store7d".net_sec
GROUP BY *, time(1h)
END
```

In the example, I process two fields, but I'm searching for a more general solution that processes all fields, a regex match or the like. With kapacitor, I don't see a solution yet; maybe splitting, processing and joining fields?

```
var val_raw = stream
|from()
.database('telegraf')
.retentionPolicy('store7d')
.measurement('net_sec')
.groupBy(*)
|default()
.field('bytes_sec_recv', 0)
.field('bytes_sec_sent', 0)

var windowed_vals = val_raw
|window()
.period(30s)
.every(30s)
.align()

var bytes_sec_recv_avg = windowed_vals
|mean('bytes_sec_recv')
.as('bytes_sec_recv_avg')

var bytes_sec_sent_avg = windowed_vals
|mean('bytes_sec_sent')
.as('bytes_sec_sent_avg')

bytes_sec_recv_avg
|union(bytes_sec_sent_avg)
|influxDBOut()
.database('telegraf')
.retentionPolicy('store1y')
.measurement('net_sec')
.precision('s')
```

Is this the way to go? I'm not sure if this scales well, as there are ~8 fields (rx and tx of bytes, packets, errors and drops) and 6 aggregation functions (min, max, avg, percentile (25, 75, 95)) per telegraf net series, leading to 48 variables...
Something like

```
|mean('bytes_sec_sent', 'bytes_sec_recv')
.as('bytes_sec_sent_avg', 'bytes_sec_recv_avg')
```

or

```
|mean('*')
.postfix('_avg')
```

would be cool.

The issue centers on Kapacitor stream nodes such as mean, percentile, union, and influxDBOut. Start by tracing these entry points and their existing behavior; done means supporting scalable aggregation across matching fields with generated aggregate names, without requiring dozens of manually defined variables.

Written by the indexing model from the issue text.