Skip to content

Latest commit

 

History

History
131 lines (89 loc) · 4.78 KB

README.md

File metadata and controls

131 lines (89 loc) · 4.78 KB
Raw

clj-lambda-utils

Clojure utilities for deploying AWS Lambda (JVM) function(s) to one or multiple regions via S3.

Note! The name of the repo used to be lein-clj-lambda when it contained only the Leiningen plugin.

Usage

You can use utilities as a plugin for Leiningen or Boot (or just as Clojure library).

Note! When installing Lambda all created resource names are logged to console.

Leiningen plugin

Put [lein-clj-lambda "0.12.1"] into the :plugins vector of your project.clj (or your profile if you prefer that).

Create S3 bucket and create following configuration into project.clj

:lambda {"test" [{:api-gateway {:name "DemoApiTest"} ; Optional, if you want to access via API Gateway
                  :handler "lambda-demo.LambdaFn"
                  :memory-size 512
                  :timeout 60
                  :function-name "my-func-test"
                  :environment {"MY_ENVIRONMENT_VAR" "some value" ;Optional
                                "SOME_OTHER_ENV_VAR" "another val"}
                  :region "eu-west-1" ; Optional, when not specified the default region specified in your AWS config will be used
                  :policy-statements [{:Effect "Allow"
                                       :Action ["sqs:*"]
                                       :Resource ["arn:aws:sqs:eu-west-1:*"]}]
                  :s3 {:bucket "your-bucket"  ; Optional, if not specified default bucket will be generated
                       :object-key "lambda.jar"}
                  :vpc {:security-group-ids ["sg-xxxx"]
                        :subnet-ids ["subnet-xxxxxx"]}}]
          "production" [{:api-gateway {:name "DemoApiProduction"} ; Optional, if you want to access via API Gateway
                         :handler "lambda-demo.LambdaFn"
                         :memory-size 1024
                         :timeout 300
                         :function-name "my-func-prod"
                         :environment {"MY_ENVIRONMENT_VAR" "some value"
                                       "SOME_OTHER_ENV_VAR" "another val"}
                         :region "eu-west-1" ; Optional, when not specified the default region specified in your AWS config will be used
                         :s3 {:bucket "your-bucket"
                              :object-key "lambda.jar"}
                         :vpc {:security-group-ids ["sg-xxxx"]
                               :subnet-ids ["subnet-xxxxxx"]}}]}

Then run

$ lein lambda install test

or

$ lein lambda install production

This will create S3 buckets that will be used for uploading code to Lambda. Also, this creates new IAM roles and policies so the Lambda function can write to Cloudwatch logs. If you need to set up additional access rights, you can pass :policy-statements. The format of statements are specified in a Clojure EDN map but they will be passed as JSON to AWS IAM (See here the details of policy statements http://docs.aws.amazon.com/IAM/latest/UserGuide/reference_policies_elements.html#Statement) If an S3 bucket or role already exists, its creation will be skipped.

If the Lambda function already exists and you want to just configure API gateway for that you can run:

$ lein lambda install test --only-api-gateway

After the Lambda function is installed you should not run install anymore but instead just run the update task to update the latest code to Lambda environment.

$ lein lambda update test

or

$ lein lambda update production

You can delete Lambda and all related resources with following command

$ lein lambda uninstall test

Boot task

A single boot task lambda is provided.

Put [boot-clj-lambda "0.2.0"] into your boot dependency vector.

(require '[boot-clj-lambda.boot :refer [lambda]])

(boot (lambda :action           :update           ;; either :update or :install
              :lambda-config    lambda-config
              :stage-name       "test"
              :jar-file         "/path/to/jar.jar"
              :only-api-gateway true))

Where lambda-config is a map just like the example in the Leiningen plug-in section above, keyed by stage-name.

It is recommended to read in lambda-config from a file before passing it to the task CLI:

$ boot -a install :lambda-config "$(< /path/to/config.edn)" ...

Library

Add the following to your project.clj :dependencies:

[clj-lambda "0.8.1"]

Then run

(require '[clj-lambda.aws :as aws])

(aws/update-lambda stage-name config jar-file opts)

(aws/install-lambda stage-name config jar-file opts)

(aws/uninstall-lambda stage-name config opts)

where config is a vector of configuration options; see the example from the Leiningen plugin documentation above.

License

Copyright © 2016-2017 Markus Hjort

Distributed under the Eclipse Public License 1.0.