Managing Connections

This section describes how to connect the Python Analytics SDK to an Enterprise Analytics cluster. It contains best practices as well as information on TLS/SSL and advanced connection options.

Our Getting Started pages cover the basics of making a connection to an Enterprise Analytics cluster. This page is a wider look at the topic.

Connecting to a Cluster

The examples below use these imports:

from couchbase_analytics.cluster import Cluster
from couchbase_analytics.credential import Credential
from couchbase_analytics.options import (ClusterOptions,
                                        SecurityOptions)

A connection to an Enterprise Analytics cluster is represented by a Cluster object. The simplest way to create a Cluster object is to call Cluster.create_instance() with a connection string, user credentials, and any optional settings:

def main() -> None:
    # Update this to your cluster
    connstr = 'https://--your-instance--'
    username = 'username'
    pw = 'Password!123'
    # User Input ends here.

    cred = Credential.from_username_and_password(username, pw)
    cluster = Cluster.create_instance(connstr, cred, opts)

Connection Strings

Typically, an Enterprise Analytics cluster will be behind a load balancer, and you will be making a connection over TLS — so the port used will be 443. This is the defaut for the SDK, so port 443 does not need to be specified:

https://analytics.example.com

You must specify the schema — either https:// (for TLS) or http:// (for insecure connections — perhaps on a development machine) in the connection string. The default port for insecure connections is port 80.

If you’re connecting to a cluster directly, without a load balancer, you can specify the port in the connection string:

https://analytics.example.com:18095

For a standalone Analytics cluster, the port is usually 18095 (or 8095 for an insecure connection). Make sure to check with your administrator.

Client Settings Parameters

Connection strings can also include client settings, which will override any that are also set in the code.

Connection string with two parameters

https://analytics.example.com?timeout.connect_timeout=30s&timeout.query_timeout=2m

The full list of recognized parameters is documented in the client settings reference.

Authentication by Credential

Similarly to the Authenticator abstraction in Couchbase Operational SDKs, Analytics SDKs use a Credential abstraction covering regular password authentication (Basic Access Authentication), JSON Web Tokens (JWT), and Client Certificates through mTLS.

Basic Access Authentication is shown in the example above.

JSON Web Tokens (JWT)

From the 1.1 SDK (with Enterprise Analytics Server 2.2+) JWT is supported.

def main() -> None:
    _EXAMPLE_JWT = 'eyJhbGbiOiJSUzI1NiIsInR5cCI6IkpXVCJ9.payload.signature'
    endpoint = 'https://--your-instance--'
    cred = Credential.from_jwt(_EXAMPLE_JWT)
    cluster = Cluster.create_instance(endpoint, cred)

Certificate Authentication

From the 1.1 Analytics SDK (with Enterprise Analytics Server 2.2+) certificate authentication is supported. A conceptual and architectural overview of Enterprise Analytics’s support of X.509 certificates is provided in the Server certificates docs. Practical information on handling certificates can be found in the Enterprise Analytics certificates docs.

The Analytics SDK authenticates the client during the TLS handshake. The SDK reads the certificate and private key from a PKCS#12 file:

clientdir = "etc/x509-cert/SSLCA/clientdir"
endpoint = 'https://--your-instance--'
cred = Credential.from_certificate(os.path.join(clientdir, "client.pem"),
                                   os.path.join(clientdir, "client.key"))
cluster = Cluster.create_instance(endpoint, cred)

Certificate Authority

To make a TLS connection to an Enterprise Analytics cluster with a root certificate issued by a trusted CA (Certificate Authority), you do not need to add this to your configuration — the platform’s defaults are automatically trusted.

The cluster’s root certificate just needs to be issued by a CA whose certificate is in your system trust store. This includes well known CAs (including GoDaddy and Verisign), plus any other CA certificates that you wish to add.

Local Development

We strongly recommend that the client and server are in the same LAN-like environment (e.g. AWS Region).