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.
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.
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
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)" ...
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.
Copyright © 2016-2017 Markus Hjort
Distributed under the Eclipse Public License 1.0.