For support requests please search or create posts on Serverless Forum: https://forum.serverless.com
I'm reading & assigning values from a config.yml file, which uses booleans to indicate whether a function should be deployed or not. From there, I am using the serverless-plugin-ifelse plugin to conditionally exclude functions from my serverless.yml. The issue is that apparently it requires a string on the left hand condition when referencing a variable.
Getting error - "Trying to populate non string value into a string for variable ${strToBool(false)}"
I have to use the strToBool utility because sometimes the boolean property isnt always set in the config.yml (indicating the function should not be deployed).
custom:
config: ${file(${opt:configFile})}
myLambdaFunction: ${strToBool(${self:custom.config.functions.myLambdaFunction.active, 'false'})}
serverlessIfElse:
- If: '"${self:custom.myLambdaFunction}" == "false"'
Exclude:
- functions.myLambdaFunction
serverless deploy works just fine
Hello, everyone! Just wanted to get expert opinion.
Problem: I have interdependent processes. What would be the appropriate way to architect it? Series of lambda functions invoking each other through SQS? For example, function 1 invoked by SQS1 and pushed data to SQS2, which invokes function 2 and so on.
SQS queue 1 --> Invokes Function 1 --> output goes to SQS queue 2 --> Invokes Function 2 --> SQS queue 3 ....
Hey, I don’t know if this is the best place to ask this but I’m looking at making a contribution to serverless with a few package updates to clear up a few security updates to various sub-dependencies. One of them would involve an update to the “boxen” dependency from 3.2.0 to the next version which is 4.x. In the 4.x update there was a change that made it require node 8+.
I can see from the travis build the project is still testing on node 6. Are there plans to remove node 6 support at any time? Would the project be open to some package updates that break node 6 support to get rid of some security issues in sub dependencies? I only ask as node 6 is out of LTS support with the node project now.
Foo:
handler: handler.hello
events:
- s3:
bucket:demo-bucket
event: s3:ObjectCreated:*
rules:
- prefix: filesPng/
- suffix: .png
existing: true
...
resources:
Resources:
FooLambdaPermission:
Type: "AWS::Lambda::Permission"
Properties:
Action: lambda:InvokeFunction
FunctionName:
'Fn::GetAtt':
- FooLambdaFunction
- Arn
Principal: "s3.amazonaws.com"
SourceArn: arn:aws:s3:::demo-bucket
SourceAccount:
Ref: "AWS::AccountId"
Hello @vviswaroop , if you are handling all in one .yml file and wants to speed up the deployment time, what you can do is to deploy "only the function that has new changes" as you can see in the Framework documentation
https://serverless.com/framework/docs/providers/aws/guide/deploying/#deploy-function
Hi all!! I'm working on a multi-stack deployment. One stack will have Cognito and some basic authentication endpoints to retrieve a token, and I'm exporting Cognito, as you can see in the image. In the other stacks I'm trying to import that like this:
CognitoUserPool:
Fn::ImportValue:
Fn::Sub "${cf:CAS-API-AUTH${opt:service-suffix,''}-${self:provider.stage}.CognitoUserPool}"When I go to deploy, I see the following:
$ sls deploy -s dev --service-suffix Larry
Serverless Error ---------------------------------------
Trying to request a non exported variable from CloudFormation. Stack name: "CAS-API-AUTHLarry-dev" Requested variable: "CognitoUserPool".
It looks like it's trying to import the variable that's there, but it can't find it. I'm not sure what I'm doing wrong. Thanks in advance!!
@iamrahulroy I guess it depends on how often these processes trigger, if its sort of infrequent then invoke-ing other lambdas may be fine, but if they trigger frequently you could profit from batch processing by using a SQS queue
Yes, it's frequent and that is why we choose SQS.
I've been using the serverless framework for some time now. With it I use the serverless wsgi to invoke a serverless flask application for handling my API. Recently I started a new project following the same pattern but this time the API has a cognito authorizer. When working with this I was having a hard time getting cognito to work unless I added the integration attribute as lambda inside of the HTTP event.
functions:
api:
handler: wsgi_handler.handler
events:
- http:
path: /api
method: ANY
cors: true
integration: lambda
authorizer:
type: COGNITO_USER_POOLS
authorizerId:
Ref: ApiGatewayAuthorizer
- http:
path: /api/{proxy+}
method: ANY
cors: true
integration: lambda
authorizer:
type: COGNITO_USER_POOLS
authorizerId:
Ref: ApiGatewayAuthorizerAt that point I thought I was golden but I realized that really I was getting a new error message returned from The serverless_wsgi.py. Digging into this file what I have learned was that the path info that's attached to the event seems to be different than what was expected. Resulting in the error message that kills the whole thing.
# If a user is using a custom domain on API Gateway, they may have a base
# path in their URL. This allows us to strip it out via an optional
# environment variable.
path_info = event[u"path"]
base_path = os.environ.get("API_GATEWAY_BASE_PATH")
if base_path:
script_name = "/" + base_path
if path_info.startswith(script_name): # ERROR HERE: 'dict' object has no attribute 'startswith'
path_info = path_info[len(script_name) :]path info in this scenario is {'proxy': 'devices'} instead of /path/to/resource
So my question to the room is that I'm curious if anyone has done this or if I'm on an edge case? Or have I just overlooked something? I'm not sure what information to provide but I am using a custom domain and base path.
Framework Core: 1.61.3