Claudia.js 1.9.0

• By Gojko Adzic
• Sep 21, 2016

claudia 1.9.0 is now available from NPM. Here are the changes:

Reuse code easier across different functions

Previously, to reuse code, you either had to manually manage local dependencies or publish a NPM module. Claudia can now work working with local relative dependencies in package.json (referencing relative directories on your disk). All the formats that NPM supports for local references are supported.

Deploy from storage-restricted environments

Claudia now cleans up after itself better, removing temporary files produced for packaging, so you can use it easier in situations where you don’t control the temporary file storage, such as cloud continuous integration systems. (Thanks to Philipp Holly)

Specify --keep with claudia create or claudia update to keep the zip files around for troubleshooting (claudia will print out the location of the archive in that case).

Audit uploads easier, and deploy larger functions more reliably

You can now use S3 buckets as an intermediary for installing code to Lambda. Supply a S3 bucket name with --use-s3-bucket <bucket-name> when using claudia create or claudia update and Claudia will send a binary archive to S3 first, then install it to Lambda from there, instead of uploading code directly to Lambda. Claudia will also print out the uploaded file key in the command results, so you can easily integrate it with auditing tools.

Using an intermediate S3 bucket to upload helps with larger packages, and especially over low-bandwidth connections. Lambda has a timeout of two minutes for direct code uploads.

Claudia will not clean up the S3 files after install, but you can easily set up a S3 bucket to automatically delete files older than a few days.