Lambda Specific Configuration does not work with CDK

[hnishi]

I am using Chalice with CDK to create a REST API and a scheduled lambda function (for a batch job).

I want to specify the configuration for the scheduled lambda function.

@app.schedule(Cron(0, 16, "L", "*", "?", "*"))
def foo(event, context):
    print("hello world")

In the CDK file, the following configuration is used.

class ChaliceApp(cdk.Stack):
    def __init__(self, scope, id, **kwargs):
        super().__init__(scope, id, **kwargs)
        self.dynamodb_table = self._create_ddb_table()
        self.chalice = Chalice(
            self,
            "ChaliceApp",
            source_dir=RUNTIME_SOURCE_DIR,
            stage_config={
                "environment_variables": {
                    "APP_TABLE_NAME": self.dynamodb_table.table_name
                },
                "lambda_functions": {
                    "foo": {
                        "lambda_timeout": 600,
                        "lambda_memory_size": 256,
                        "environment_variables": {"HELLO": "world"},
                    }
                },
            },
        )
        self.dynamodb_table.grant_read_write_data(self.chalice.get_role("DefaultRole"))

However, after deploying the application with cdk deploy, the scheduled lambda function is created with the default timeout (60 sec) and memory size (128 MB).

I think the Lambda Specific Configuration does not work with CDK.

[hnishi]

The current code only reads _config_from_disk in config.py.

chalice/chalice/config.py

Lines 197 to 209 in f64bd82

	if varies_per_function:
	# search order:
	# config['stages']['lambda_functions']
	# config['stages']
	# config['lambda_functions']
	search_dicts.insert(
	0, self._config_from_disk.get('stages', {}).get(
	self.chalice_stage, {}).get('lambda_functions', {}).get(
	self.function_name, {}))
	search_dicts.append(
	self._config_from_disk.get('lambda_functions', {}).get(
	self.function_name, {}))
	search_dicts.extend([self._config_from_disk, self._default_params])

I created a PR.

#2000

I found a workaround to set up the lambda specific configuration with CDK and the current code.

Add the following configuration in ../runtime/.chalice/config.json.

In the CDK configuration step, the stage name will be the app_name (e.g. cdkdemo in this example). Thus add the configuration for the stage name.

{
  "version": "2.0",
  "app_name": "cdkdemo",
  "stages": {
      "cdkdemo": {
            "lambda_functions": {
                "foo": {
                    "lambda_timeout": 200,
                    "lambda_memory_size": 256,
                    "environment_variables": {"HELLO": "world2"}
                }
            }
    },
    "dev": {
      "api_gateway_stage": "api",
      "lambda_functions": {
        "api_handler": {
          "environment_variables": {
            "APP_TABLE_NAME": "xxx"
          }
        },
        "foo": {
          "environment_variables": {
            "APP_TABLE_NAME": "yyy"
          },
          "lambda_timeout": 900,
          "lambda_memory_size": 256
        }
      }
    }
  }
}

[MacHu-GWU]

I feel like the chalice + cdk deployment is super inmature. lots of things are not configurable and the old .chalice/config.json is not working and also not mentioned in the document. Don't get me wrong, I am a big fan of chalice.

I am working a a team having around 250+ lambda function in production and 10+ different internal api. We have tried pure CDK, Chalice + CDK, and eventually we choose CDK for iam role, and infra, then explicitly use those values in Chalice / config.json and deploy in the "classic chalice" way, and use S3 to store deployed.json

[mateuspadua]

I think this answer can help:
#1846 (comment)