Skip to main content

PostgreSQL utility library for AWS Lambda

Project description

Soong - PostgreSQL utility library for AWS Lambda

The purpose of this library is to let Lambda functions connect to PostgreSQL on RDS and run queries against it.

⚠️ Soong is currently in active development and should be considered work in progress. Do not use in production! ⚠️

How working with a relational DB on Lambda is different

  1. Authentication
  2. Connection Reuse
  3. Resource Constraints


Non-Secret Parameters

Database connection parameters, such as the host, port, database name and user name, should be passed to your lambda functions using environment variables. Soong supports this pattern by automatically detecting those variables.

If you are using SAM, you can assign those parameters to all of your functions in the global section of the template, as shown below. Soong will pick these variables automatically.

        PG_HOST: ""
        PG_DBNAME: database1
        PG_USER: jane_doe


A database password however, should not be used at all! Instead, use IAM authentication, as explained in the RDS user guide.

Soong supports this authentication pattern by automatically generating an auth token when running in the Lambda environment.

In order to enable this pattern, you must:

  1. Enable IAM db authentication for your RDS instance (user guide).
  2. Create a database user that uses IAM authentication (user guide). You can also GRANT rds_iam to an existing user.
  3. Create an IAM policy for IAM database access (user guide). Make sure that the policy covers the user from the previous step.
  4. Assign that policy to the role under which your functions execute (user guide).

Finally, you must use RDS API to generate an auth-token to use in place of a password. Soong does that for you behind the scenes whenever you create a connection.

Code Example

If you did all of the above, connecting to the database from your code is now a breeze:

import soong
# The fancy cursor_factory is optional
with soong.connect(cursor_factory=psycopg2.extras.NamedTupleCursor) as conn:
    with conn.cursor() as cursor:
        cursor.execute("SELECT 'Hello, world!' AS hello;")
        assert cursor.fetchone().hello == 'Hello, world!'

Connection Reuse

In server-full applications, it is a best practice to manage a connection pool that lets various modules in the application reuse database connections. When you decompose an application into serverless-style functions, there is usually no point in a connection pool, because every function most likely needs just one connection. This is why Soong makes it simple to connect, but does not provide a connection pool.

This does not mean that there is no connection reuse. Code that runs outside of the handler function gets reused if the container remains warm and is reused by another invocation of the function.

Moreover, Soong provides a global connection and makes it simple to initialize it lazily, and reconnect when it's closed. See the code examples below.

Code Examples

# A global connection object is defined inside the soong package
import soong

def lambda_handler(event, context):
    with soong.connection() as conn:  # Lazy initialization
        # Run some queries inside a transaction

You can also control the connection yourself, rather than use Soong's. The code example below shows how to generate the connection at the module scope, outside of the handler, so it can be reused:

import soong

conn = soong.connect()

def lambda_handler(event, context):
    with conn:
        # Run some queries inside a transaction

Note that you can also generate connections inside the handler if you want them to close rather than remain open for reuse. There are some scenarios where this is the better option, such as when your function rarely gets called, so there is no connection reuse in practice, but every invocation keeps a connection hanging for a while, wasting resources on your database server.

Resource Constraints

With Lambda you pay directly for the CPU/Memory and execution time. Therefore it is important to limit resource utilization to the minimum necessary. In addition, code in lambda functions tends to be short and simple.

In my opinion, with most lambda functions it makes sense to avoid ORM tools, such as SQL Alchemy, and opt for a more direct approach that is also more lightweight.

The Soong query and dml modules provide some well-documented utility functions that save you from writing a lot of boilerplate code in the most common database access use cases.

Code Example

import soong

# Get the gizmo with id 42 from the gizmos table
gizmo42 = soong.query.get(soong.connection(cursor_factory=psycopg2.extras.NamedTupleCursor), 'gizmos', 42)
print(f'Got gizmo {}')

# Now rename it
# Re-uses the global connection object, that's already been initialized with the cursor_factory
soong.dml.update(soong.connection(), 'gizmos', 42, {'name': 'Gizmo42: The Next Generation'})

# Add a new gizmo
id = soong.dml.insert(soong.connection(),
    'gizmos', {'name': 'Best gizmo ever!', 'color': 'Bright red'}, returning='id')

# Run a SELECT
for gizmo in, 'gizmos', {'color': 'pink'}):

# You can also run an arbitrary query, without the boilerplate
for gizmo in soong.query.execute(soong.connection(), 'SELECT * FROM gizmos WHERE color = %s', ('pink', )):

Why "Soong"?

The name is a reference to Noonien Soong, the human cyberneticist who created Data, the android character in Star Trek: The Next Generation.

A hologram of Dr. Soong from around the time of Data's creation


Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for soong, version 0.6.0
Filename, size File type Python version Upload date Hashes
Filename, size soong-0.6.0-py3-none-any.whl (9.9 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size soong-0.6.0.tar.gz (13.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page