Here's my client:
https://gist.github.com/tschm/4aa4eeda016a326331fbc024c0b12454

5.2.0 was released a day after you reported this, try that one. Also try 5.0.0 and see if it is equally slow.

We also experience massive performance problems on writing large-ish data frames with version 5.1.0. One example (see below) uses 5 seconds on my local laptop to write a million lines (one column) with version 5.0.0, but on version 5.1.0 it takes minutes (in fact I gave up before it finished). On the current master the performance is back to 5.0.0 performance. The pypi current latest version 5.2.0 is broken (see #616), so I'll use 5.0.0 or master for now. I guess the fix in #617 will come to the pip-world soon though.

I have no idea what the cause is, maybe something wrong with our setup of InfluxDB itself perhaps...

The below example produces this output for master and similar for 5.0.0:
InfluxDB python version: 5.0.0
Number of points: 10000
Time: 0.20903187998919748
Number of points: 100000
Time: 0.6698060079943389
Number of points: 1000000
Time: 5.531810616987059

But for version 5.1.0 this is how it looks:
InfluxDB python version: 5.1.0
Number of points: 10000
Time: 3.2126710579905193
Number of points: 100000
Time: 34.53647285097395
Number of points: 1000000
<Aborted here, didn't bother to wait>

Example code used (Tried on both 1.5.X and 1.6.0 InfluxDB instance):

from influxdb import DataFrameClient
import pandas as pd
from timeit import default_timer as timer

host = '0.0.0.0'
port = 8086
user = 'admin'
password = ''
db_name = 'test'

def simple_test(num_points, batch_size):
    client = DataFrameClient(host, port, user, password, db_name)
    x = pd.DatetimeIndex(start=pd.Timestamp("2010-01-01"), periods=num_points, freq="S")
    y = pd.DataFrame(index=x, data=pd.np.random.randn(num_points, 1))

    client.create_database(db_name)
    client.write_points(y, "perf-test", batch_size=batch_size, protocol='line')
    client.drop_database(db_name)

if __name__=='__main__':
    print("InfluxDB python version: {}".format(influxdb.__version__))
    num_points_list = [10000, 100000, 1000000]
    batch_size = 10000
    for num_points in num_points_list:
        print("Number of points: {}".format(num_points))
        start = timer()
        simple_test(num_points, batch_size)
        end = timer()
        elapsed = end - start
        print("Time: {}".format(elapsed))