High CPU consumption with InfluxDB 1.7.9

Hi,

We're using InfluxDB with telegraf on a monitoring server. Only telegraf database is present in InfluxDB.
With telegraf, we use the 'tail' plugin to insert data from a file using InfluxDB format. All data is stored in measurement "loopingd"
We have a script to check data present in influxDB. This script run the following query very regularly:

```
select count(rtt), mean(rtt) from loopingd where server='$server' and time >= '$start_ts' and time < '$end_ts' group by time(1d)
```

Where:
- `$server` is the server we want to check
- `$start_ts` is a timestamp in nanoseconds corresponding to previous minute, rounded to the minute (exemple: if now is 15:30:34, this will be the timestamp for 15:29:00).
- `$end_ts`is a timestamp in nanoseconds corresponding to current minute, rounded to the minute (exemple: if now is 15:30:34, this will be the timestamp for 15:30:00).

We currently have 24 097 076 points stored in this measurement.

After upgrading InfluxDB to v1.7.9 we observed an increase in CPU usage from 15% usage to 53% usage.
When doing a `top` command, we saw a lot of cpu consummed by the queries above.

__Steps to reproduce:__
I'm not sure on how to reproduce the issue, so I will try to explain how to create a setup like mine.
1. Install InfluxDB 1.7.8
2. Create a database named telegraf
3. Insert 50 entries in a measurement `loopingd` with following labes: `host`, `path`, `server` and a field named `rtt`
4. Perform multiple queries `select count(rtt), mean(rtt) from loopingd where server='$server' and time >= '$start_ts' and time < '$end_ts' group by time(1d)` and measure cpu usage on server
5. Upgrade to InfluxDB 1.7.9
6. Execute the same queries and measure CPU usage

__Expected behavior:__
CPU usage is the same between point 4 and 6

__Actual behavior:__
CPU usage is higher at point 6 than at point 4.

__Environment info:__

* System info: Run `uname -srm` and copy the output here
```
Linux 4.4.0-1096-aws x86_64
```
* InfluxDB version: Run `influxd version` and copy the output here
```
InfluxDB v1.7.9 (git: 1.7 23bc63d43a8dc05f53afa46e3526ebb5578f3d88)
```
* Other relevant environment details: Container runtime, disk info, etc
```
N/A
```

__Config:__
Copy any non-default config values here or attach the full config as a gist or file.

We use the default configuration file.

__Logs:__
Include snippet of errors in log.

None provided

__Performance:__
Generate profiles with the following commands for bugs related to performance, locking, out of memory (OOM), etc.

```sh
# Commands should be run when the bug is actively.
# Note: This command will run for at least 30 seconds.
curl -o profiles.tar.gz "http://localhost:8086/debug/pprof/all?cpu=true"
curl -o vars.txt "http://localhost:8086/debug/vars"
iostat -xd 1 30 > iostat.txt

# Attach the `profiles.tar.gz`, `vars.txt`, and `iostat.txt` output files.
```
[profiles.tar.gz](https://github.com/influxdata/influxdb/files/3784165/profiles.tar.gz)
[vars.txt](https://github.com/influxdata/influxdb/files/3784168/vars.txt)
[iostat.txt](https://github.com/influxdata/influxdb/files/3784178/iostat.txt)

Start by comparing the repeated loopingd queries on InfluxDB 1.7.8 and 1.7.9 using the reproduction steps and the attached profiles.tar.gz, vars.txt, and iostat.txt. Investigate the CPU difference for the time-bounded aggregate query filtered by server; done means identifying and correcting the 1.7.9 regression so CPU usage matches the earlier version.

Written by the indexing model from the issue text.