Google Data studio uses PostgreSQL connector with Heroku Postgres

If you want to create Data Source by PostgreSQL from Heroku Postgresql you need to enable SSL when connecting to the database.
  • I found a solution 😃
1. Create client key and client certificate by OpenSSL.
	openssl req \
       -newkey rsa:2048 -nodes -keyout client.key \
       -x509 -days 365 -out client.crt
2. Use the python script to create certificate file.
  • Need to run this python script to create a certificate file.
#!/usr/bin/env python

import argparse
import socket
import ssl
import struct
import subprocess
import sys

    from urlparse import urlparse
except ImportError:
    from urllib.parse import urlparse

def main():
    args = get_args()
    target = get_target_address_from_args(args)
    sock = socket.create_connection(target)
        certificate_as_pem = get_certificate_from_socket(sock)
    except Exception as exc:
        sys.stderr.write('Something failed while fetching certificate: {0}\n'.format(exc))

def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument('database', help='Either an IP address, hostname or URL with host and port')
    return parser.parse_args()

def get_target_address_from_args(args):
    specified_target = args.database
    if '//' not in specified_target:
        specified_target = '//' + specified_target
    parsed = urlparse(specified_target)
    return (parsed.hostname, parsed.port or 5432)

def get_certificate_from_socket(sock):
    ssl_context = get_ssl_context()
    sock = ssl_context.wrap_socket(sock)
    certificate_as_der = sock.getpeercert(binary_form=True)
    certificate_as_pem = encode_der_as_pem(certificate_as_der)
    return certificate_as_pem

def request_ssl(sock):
    # 1234.5679 is the magic protocol version used to request TLS, defined
    # in pgcomm.h)
    version_ssl = postgres_protocol_version_to_binary(1234, 5679)
    length = struct.pack('!I', 8)
    packet = length + version_ssl

    data = read_n_bytes_from_socket(sock, 1)
    if data != b'S':
        raise Exception('Backend does not support TLS')

def get_ssl_context():
    # Return the strongest SSL context available locally
    for proto in ('PROTOCOL_TLSv1_2', 'PROTOCOL_TLSv1', 'PROTOCOL_SSLv23'):
        protocol = getattr(ssl, proto, None)
        if protocol:
    return ssl.SSLContext(protocol)

def encode_der_as_pem(cert):
    # Forking out to openssl to not have to add any dependencies to script,
    # preferably you'd do this with pycrypto or other ssl libraries.
    cmd = ['openssl', 'x509', '-inform', 'DER']
    pipe = subprocess.PIPE
    process = subprocess.Popen(cmd, stdin=pipe, stdout=pipe, stderr=pipe)
    stdout, stderr = process.communicate(cert)
    if stderr:
        raise Exception('OpenSSL error when converting cert to PEM: {0}'.format(stderr))
    return stdout.strip()

def read_n_bytes_from_socket(sock, n):
    buf = bytearray(n)
    view = memoryview(buf)
    while n:
        nbytes = sock.recv_into(view, n)
        view = view[nbytes:] # slicing views is cheap
        n -= nbytes
    return buf

def postgres_protocol_version_to_binary(major, minor):
    return struct.pack('!I', major << 16 | minor)

if __name__ == '__main__':
# Enable your python environment and run this script.
# Run python script follows this example command.
python postgres_get_server_cert {heroku_db_hostname}:{port} > server_certificate.crt
notion image
3. Fill all these required parameters to google data studio then click AUTHENTICATE.