Why does count function take 1000 seconds with a 70 million rows of data? How can I solve it?

Question

cnosdb version 2.3.

I used SQL command

select count(*) from readings;

to do a simple aggregate execution. And I got this result after waiting for more than 1000 seconds:

+-----------------+
| COUNT(UInt8(1)) |
+-----------------+
| 77784797        |
+-----------------+
Query took 1113.768 seconds.

How can I solve it? It takes too much time.