Product docs and API reference are now on Akamai TechDocs.
Search product docs.
Search for ββ in product docs.
Search API reference.
Search for ββ in API reference.
Search Results
 results matching 
 results
No Results
Filters
Migrating from AWS Lambda to Knative
Traducciones al EspaΓ±olEstamos traduciendo nuestros guΓas y tutoriales al EspaΓ±ol. Es posible que usted estΓ© viendo una traducciΓ³n generada automΓ‘ticamente. Estamos trabajando con traductores profesionales para verificar las traducciones de nuestro sitio web. Este proyecto es un trabajo en curso.
Knative is an open source platform that extends Kubernetes to manage serverless workloads. It provides tools to deploy, run, and manage serverless applications and functions, enabling automatic scaling and efficient resource usage. Knative consists of several components:
- Serving: Deploys and runs serverless containers.
- Eventing: Facilitates event-driven architectures.
- Functions: Deploys and runs functions locally and on Kubernetes.
This guide walks through the process of migrating an AWS Lambda function to a Knative function running on the Linode Kubernetes Engine (LKE).
Before You Begin
Read our Getting Started with Linode guide, and create a Linode account if you do not already have one.
Create a personal access token using the instructions in our Manage personal access tokens guide.
Ensure that you have Git installed.
Follow the steps in the Install kubectl section of our Getting started with LKE guide to install
kubectl
.Install the Linode CLI using the instructions in our Install and configure the CLI guide.
Ensure that you have Knative’s
func
CLI installed.Ensure that you have Docker installed and have a Docker Hub account.
Install
jq
, a lightweight command line JSON processor:sudo apt install jq
sudo
. If youβre not familiar with the sudo
command, see the Users and Groups
guide.Provision a Kubernetes Cluster
While there are several ways to create a Kubernetes cluster on Linode, this guide uses the Linode CLI to provision resources.
Use the Linode CLI command (
linode
) to see available Kubernetes versions:linode lke versions-list
ββββββββ β id β ββββββββ€ β 1.31 β ββββββββ€ β 1.30 β ββββββββ€ β 1.29 β ββββββββ
It’s generally recommended to provision the latest version of Kubernetes unless specific requirements dictate otherwise.
Use the following command to list available Linode plans, including plan ID, pricing, and performance details. For more detailed pricing information, see Akamai Connected Cloud: Pricing :
linode linodes types
The examples in this guide use the g6-standard-2 Linode, which features two CPU cores and 4 GB of memory. Run the following command to display detailed information in JSON for this Linode plan:
linode linodes types --label "Linode 4GB" --json --pretty
[ { "addons": { "backups": { "price": { "hourly": 0.008, "monthly": 5.0 }, "region_prices": [ { "hourly": 0.009, "id": "id-cgk", "monthly": 6.0 }, { "hourly": 0.01, "id": "br-gru", "monthly": 7.0 } ] } }, "class": "standard", "disk": 81920, "gpus": 0, "id": "g6-standard-2", "label": "Linode 4GB", "memory": 4096, "network_out": 4000, "price": { "hourly": 0.036, "monthly": 24.0 }, "region_prices": [ { "hourly": 0.043, "id": "id-cgk", "monthly": 28.8 }, { "hourly": 0.05, "id": "br-gru", "monthly": 33.6 } ], "successor": null, "transfer": 4000, "vcpus": 2 } ]
View available regions with the
regions list
command:linode regions list
With a Kubernetes version and Linode type selected, use the following command to create a cluster named
knative-playground
in theus-mia
(Miami, FL) region with three nodes and auto-scaling. Replace knative-playground and us-mia with a cluster label and region of your choosing, respectively:linode lke cluster-create \ --label knative-playground \ --k8s_version 1.31 \ --region us-mia \ --node_pools '[{ "type": "g6-standard-2", "count": 3, "autoscaler": { "enabled": true, "min": 3, "max": 8 } }]'
Once your cluster is successfully created, you should see output similar to the following:
Using default values: {}; use the --no-defaults flag to disable defaults ββββββββββββββββββββββ¬βββββββββ¬ββββββββββββββ β label β region β k8s_version β ββββββββββββββββββββββΌβββββββββΌββββββββββββββ€ β knative-playground β us-mia β 1.31 β ββββββββββββββββββββββ΄βββββββββ΄ββββββββββββββ
Access the Kubernetes Cluster
To access your cluster, fetch the cluster credentials in the form of a kubeconfig
file.
Use the following command to retrieve the cluster’s ID:
CLUSTER_ID=$(linode lke clusters-list --json | \ jq -r \ '.[] | select(.label == "knative-playground") | .id')
Create a hidden
.kube
folder in your user’s home directory:mkdir ~/.kube
Retrieve the
kubeconfig
file and save it to~/.kube/lke-config
:linode lke kubeconfig-view --json "$CLUSTER_ID" | \ jq -r '.[0].kubeconfig' | \ base64 --decode > ~/.kube/lke-config
Once you have the
kubeconfig
file saved, access your cluster by usingkubectl
and specifying the file:kubectl get no --kubeconfig ~/.kube/lke-config
NAME STATUS ROLES AGE VERSION lke242177-380780-1261b5670000 Ready <none> 49s v1.31.0 lke242177-380780-3496ef070000 Ready <none> 47s v1.31.0 lke242177-380780-53e2290c0000 Ready <none> 51s v1.31.0
Note Optionally, to avoid specifying
--kubeconfig ~/.kube/lke-config
with everykubectl
command, you can set an environment variable for your current terminal session:export KUBECONFIG=~/.kube/lke-config
Then run:
kubectl get no
Set Up Knative on LKE
There are multiple ways to install Knative on a Kubernetes cluster . The examples in this guide use the YAML manifests method.
Run the following command to install the Knative CRDs:
RELEASE=releases/download/knative-v1.15.2/serving-crds.yaml kubectl apply -f "https://github.com/knative/serving/$RELEASE"
Upon successful execution, you should see a similar output indicating that the CRDs are configured:
customresourcedefinition.apiextensions.k8s.io/certificates.networking.internal.knative.dev created customresourcedefinition.apiextensions.k8s.io/configurations.serving.knative.dev created customresourcedefinition.apiextensions.k8s.io/clusterdomainclaims.networking.internal.knative.dev created customresourcedefinition.apiextensions.k8s.io/domainmappings.serving.knative.dev created customresourcedefinition.apiextensions.k8s.io/ingresses.networking.internal.knative.dev created customresourcedefinition.apiextensions.k8s.io/metrics.autoscaling.internal.knative.dev created customresourcedefinition.apiextensions.k8s.io/podautoscalers.autoscaling.internal.knative.dev created customresourcedefinition.apiextensions.k8s.io/revisions.serving.knative.dev created customresourcedefinition.apiextensions.k8s.io/routes.serving.knative.dev created customresourcedefinition.apiextensions.k8s.io/serverlessservices.networking.internal.knative.dev created customresourcedefinition.apiextensions.k8s.io/services.serving.knative.dev created customresourcedefinition.apiextensions.k8s.io/images.caching.internal.knative.dev created
Next, install the Knative Serving component:
RELEASE=releases/download/knative-v1.15.2/serving-core.yaml kubectl apply -f "https://github.com/knative/serving/$RELEASE"
You should see similar output indicating that various resources are now created:
namespace/knative-serving created role.rbac.authorization.k8s.io/knative-serving-activator created clusterrole.rbac.authorization.k8s.io/knative-serving-activator-cluster created clusterrole.rbac.authorization.k8s.io/knative-serving-aggregated-addressable-resolver created clusterrole.rbac.authorization.k8s.io/knative-serving-addressable-resolver created clusterrole.rbac.authorization.k8s.io/knative-serving-namespaced-admin created clusterrole.rbac.authorization.k8s.io/knative-serving-namespaced-edit created clusterrole.rbac.authorization.k8s.io/knative-serving-namespaced-view created clusterrole.rbac.authorization.k8s.io/knative-serving-core created clusterrole.rbac.authorization.k8s.io/knative-serving-podspecable-binding created serviceaccount/controller created clusterrole.rbac.authorization.k8s.io/knative-serving-admin created clusterrolebinding.rbac.authorization.k8s.io/knative-serving-controller-admin created clusterrolebinding.rbac.authorization.k8s.io/knative-serving-controller-addressable-resolver created serviceaccount/activator created rolebinding.rbac.authorization.k8s.io/knative-serving-activator created clusterrolebinding.rbac.authorization.k8s.io/knative-serving-activator-cluster created customresourcedefinition.apiextensions.k8s.io/images.caching.internal.knative.dev unchanged certificate.networking.internal.knative.dev/routing-serving-certs created customresourcedefinition.apiextensions.k8s.io/certificates.networking.internal.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/configurations.serving.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/clusterdomainclaims.networking.internal.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/domainmappings.serving.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/ingresses.networking.internal.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/metrics.autoscaling.internal.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/podautoscalers.autoscaling.internal.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/revisions.serving.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/routes.serving.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/serverlessservices.networking.internal.knative.dev unchanged customresourcedefinition.apiextensions.k8s.io/services.serving.knative.dev unchanged image.caching.internal.knative.dev/queue-proxy created configmap/config-autoscaler created configmap/config-certmanager created configmap/config-defaults created configmap/config-deployment created configmap/config-domain created configmap/config-features created configmap/config-gc created configmap/config-leader-election created configmap/config-logging created configmap/config-network created configmap/config-observability created configmap/config-tracing created horizontalpodautoscaler.autoscaling/activator created poddisruptionbudget.policy/activator-pdb created deployment.apps/activator created service/activator-service created deployment.apps/autoscaler created service/autoscaler created deployment.apps/controller created service/controller created horizontalpodautoscaler.autoscaling/webhook created poddisruptionbudget.policy/webhook-pdb created deployment.apps/webhook created service/webhook created validatingwebhookconfiguration.admissionregistration.k8s.io/config.webhook.serving.knative.dev created mutatingwebhookconfiguration.admissionregistration.k8s.io/webhook.serving.knative.dev created validatingwebhookconfiguration.admissionregistration.k8s.io/validation.webhook.serving.knative.dev created secret/webhook-certs created
Knative relies on an underlying networking layer. Kourier is designed specifically for Knative, and the examples in this guide use Kourier for Knative networking . Use the commands below to download and install the latest Kourier release:
RELEASE=releases/download/knative-v1.15.1/kourier.yaml kubectl apply -f "https://github.com/knative-extensions/net-kourier/$RELEASE"
The output should again indicate the creation of multiple new elements:
namespace/kourier-system created configmap/kourier-bootstrap created configmap/config-kourier created serviceaccount/net-kourier created clusterrole.rbac.authorization.k8s.io/net-kourier created clusterrolebinding.rbac.authorization.k8s.io/net-kourier created deployment.apps/net-kourier-controller created service/net-kourier-controller created deployment.apps/3scale-kourier-gateway created service/kourier created service/kourier-internal created horizontalpodautoscaler.autoscaling/3scale-kourier-gateway created poddisruptionbudget.policy/3scale-kourier-gateway-pdb created
The following command configures Knative to use Kourier as the default ingress controller:
kubectl patch configmap/config-network \ --namespace knative-serving \ --type merge \ --patch \ '{"data":{"ingress-class":"kourier.ingress.networking.knative.dev"}}'
configmap/config-network patched
Note If Istio is already installed in your cluster, you may choose to reuse it for Knative as well.With Kourier configured, the Knative Serving installation now has a
LoadBalancer
service for external access. Use the following command to retrieve the external IP address in case you want to set up your own DNS later:kubectl get service kourier -n kourier-system
The output should display the external IP address of the
LoadBalancer
:NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE kourier LoadBalancer 10.128.48.124 172.235.159.7 80:31938/TCP,443:30800/TCP 4m37s
Since Kourier added several deployments, check the updated list to ensure everything is functioning correctly:
kubectl get deploy -n knative-serving
Use the output to confirm availability of the various components:
NAME READY UP-TO-DATE AVAILABLE AGE activator 1/1 1 1 7m36s autoscaler 1/1 1 1 7m36s controller 1/1 1 1 7m36s net-kourier-controller 1/1 1 1 5m7s webhook 1/1 1 1 7m36s
This guide uses the Magic DNS method to configure DNS , which leverages the sslip.io DNS service. When a request is made to a subdomain of sslip.io containing an embedded IP address, the service resolves that IP address. For example, a request to https://52.0.56.137.sslip.io returns
52.0.56.137
as the IP address. Use thedefault-domain
job to configure Knative Serving to use sslip.io:MANIFEST=knative-v1.15.2/serving-default-domain.yaml kubectl apply -f "https://github.com/knative/serving/releases/download/$MANIFEST"
Upon successful execution, you should see output confirming the creation of the
default-domain
job and service:job.batch/default-domain created service/default-domain-service created
With Knative now operational in your cluster, you can begin working with Knative Functions.
Work with Knative Functions and the func
CLI
Knative Functions is a programming model that simplifies writing distributed applications on Kubernetes and Knative. It allows developers to create stateless, event-driven functions without requiring in-depth knowledge of containers, Kubernetes, or Knative itself.
The func
CLI streamlines the developer experience by providing tools to work with Knative Functions. It allows developers to manage the entire lifecycle of functions (creating, building, deploying, and invoking). This allows for local development and testing of functions without the need for a local Kubernetes cluster.
To get started, run the following command:
func
This displays help information for managing Knative Function resources:
func is the command line interface for managing Knative Function resources Create a new Node.js function in the current directory: func create --language node myfunction Deploy the function using Docker hub to host the image: func deploy --registry docker.io/alice Learn more about Functions: https://knative.dev/docs/functions/ Learn more about Knative at: https://knative.dev Primary Commands: create Create a function describe Describe a function deploy Deploy a function delete Undeploy a function list List deployed functions subscribe Subscribe a function to events Development Commands: run Run the function locally invoke Invoke a local or remote function build Build a function container System Commands: config Configure a function languages List available function language runtimes templates List available function source templates repository Manage installed template repositories environment Display function execution environment information Other Commands: completion Output functions shell completion code version Function client version information Use "func <command> --help" for more information about a given command.
Use the following command to create an example Python function (
get-emojis
) that can be invoked via an HTTP endpoint (the default invocation method):func create -l python get-emojis
This command creates a complete directory structure with multiple files:
Created python function in /home/USERNAME/get-emojis
Examine the contents of the newly created
~/get-emojis
directory:ls -laGh get-emojis
total 48K drwxr-xr-x 3 USERNAME 4.0K Oct 9 15:57 . drwxr-x--- 9 USERNAME 4.0K Oct 9 15:57 .. -rwxr-xr-x 1 USERNAME 55 Oct 9 15:57 app.sh drwxrwxr-x 2 USERNAME 4.0K Oct 9 15:57 .func -rw-r--r-- 1 USERNAME 217 Oct 9 15:57 .funcignore -rw-r--r-- 1 USERNAME 1.8K Oct 9 15:57 func.py -rw-r--r-- 1 USERNAME 97 Oct 9 15:57 func.yaml -rw-r--r-- 1 USERNAME 235 Oct 9 15:57 .gitignore -rw-r--r-- 1 USERNAME 28 Oct 9 15:57 Procfile -rw-r--r-- 1 USERNAME 862 Oct 9 15:57 README.md -rw-r--r-- 1 USERNAME 28 Oct 9 15:57 requirements.txt -rw-r--r-- 1 USERNAME 259 Oct 9 15:57 test_func.py
While reviewing the purpose of each file is outside the scope of this guide, you should examine the
func.py
file, the default implementation that Knative generates:cat ~/get-emojis/func.py
- File: ~/get-emojis/func.py
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63
from parliament import Context from flask import Request import json # parse request body, json data or URL query parameters def payload_print(req: Request) -> str: if req.method == "POST": if req.is_json: return json.dumps(req.json) + "\n" else: # MultiDict needs some iteration ret = "{" for key in req.form.keys(): ret += '"' + key + '": "'+ req.form[key] + '", ' return ret[:-2] + "}\n" if len(ret) > 2 else "{}" elif req.method == "GET": # MultiDict needs some iteration ret = "{" for key in req.args.keys(): ret += '"' + key + '": "' + req.args[key] + '", ' return ret[:-2] + "}\n" if len(ret) > 2 else "{}" # pretty print the request to stdout instantaneously def pretty_print(req: Request) -> str: ret = str(req.method) + ' ' + str(req.url) + ' ' + str(req.host) + '\n' for (header, values) in req.headers: ret += " " + str(header) + ": " + values + '\n' if req.method == "POST": ret += "Request body:\n" ret += " " + payload_print(req) + '\n' elif req.method == "GET": ret += "URL Query String:\n" ret += " " + payload_print(req) + '\n' return ret def main(context: Context): """ Function template The context parameter contains the Flask request object and any CloudEvent received with the request. """ # Add your business logic here print("Received request") if 'request' in context.keys(): ret = pretty_print(context.request) print(ret, flush=True) return payload_print(context.request), 200 else: print("Empty request", flush=True) return "{}", 200
Note that this function acts as a server that returns the query parameters or form fields of incoming requests.
Build a Function Image
The next step is to create a container image from your function. Since the function is intended to run on a Kubernetes cluster, it must be containerized. Knative Functions facilitates this process for developers, abstracting the complexities of Docker and Dockerfiles.
Navigate into the
~/get-emojis
directory:cd ~/get-emojis
To build your function, run the following
build
command while in the~/get-emojis
directory, specifying Docker Hub (docker.io
) as the registry along with your DOCKER_HUB_USERNAME:func build --registry docker.io/DOCKER_HUB_USERNAME
This command fetches a base image and builds a Docker image from your function. You should see output similar to the following as the function image is built:
Building function image Still building Still building Yes, still building Don't give up on me Still building This is taking a while π Function built: index.docker.io/DOCKER_HUB_USERNAME/get-emojis:latest
To verify that the image is successfully created, use the following command to list your Docker images:
docker images | grep -E 'knative|get-emojis|ID'
REPOSITORY TAG IMAGE ID CREATED SIZE ghcr.io/knative/builder-jammy-base 0.4.283 204e70721072 44 years ago 1.45GB DOCKER_HUB_USERNAME/get-emojis latest IMAGE_ID 44 years ago 293MB
Note While theCREATED
timestamp may be incorrect, the image is valid.Use the
run
command to run the function locally:func run
The terminal should display output indicating that the function now runs on
localhost
at port8080
.:function up-to-date. Force rebuild with --build Running on host port 8080
With your function running, open a second terminal session and enter the following command:
curl "http://localhost:8080?a=1&b=2"
By default, this initial implementation returns the URL query parameters as a JSON object. The resulting output should be:
{"a": "1", "b": "2"}
Meanwhile, you should see the output similar to the following in your original terminal window:
Received request GET http://localhost:8080/?a=1&b=2 localhost:8080 Host: localhost:8080 User-Agent: curl/7.81.0 Accept: */* URL Query String: {"a": "1", "b": "2"}
When done, close the second terminal and stop the function in the original terminal by pressing the CTRL+C keys.
Deploy the Function
Use the
deploy
command to deploy your function to your Kubernetes cluster as a Knative function and push it to the Docker registry:func deploy
function up-to-date. Force rebuild with --build Pushing function image to the registry "index.docker.io" using the "DOCKER_HUB_USERNAME" user credentials π― Creating Triggers on the cluster β Function deployed in namespace "default" and exposed at URL: http://get-emojis.default.IP_ADDRESS.sslip.io
Once the function is deployed and the Magic DNS record is established, your Knative function is accessible through this public HTTP endpoint. The new
get-emojis
repository should also now exist on your Docker Hub account:To invoke your Knative function,
curl
the functionβs public URL, adding any required query parameters. For example:curl http://get-emojis.default.IP_ADDRESS.sslip.io/?yeah=it-works!
The output should display a JSON object containing the query parameters:
{"yeah": "it-works!"}
With your Knative function running, the next step is migrate an AWS Lambda function to Knative.
Migrate Your AWS Lambda Function to Knative
This guide examines a sample Lambda function and walks through how to migrate it to Knative. Conceptually, Lambda functions are similar to Knative functions. They both have a trigger and extract their input arguments from a context or event.
The main application logic is highlighted in the example Lambda function below:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
def handler(event, context): try: logger.info("Received event: %s", event) # The descriptions may arrive as attribute of the event descriptions = event.get("descriptions") if descriptions is None: # Parse the JSON body of the event body = json.loads(event.get("body", "{}")) logger.info("Parsed body: %s", body) descriptions = body.get("descriptions", []) logger.info("Descriptions: %s", descriptions) fuzz_emoji = FuzzEmoji() result = fuzz_emoji.get_emojis(descriptions) response = { 'statusCode': 200, 'body': json.dumps(result) } except Exception as e: response = { 'statusCode': 500, 'body': json.dumps({'error': str(e)}) } return response
This example function instantiates a FuzzEmoji
object and calls its get_emojis()
method, passing a list of emoji descriptions. The emoji descriptions may or may not map to official emoji names like fire
(π₯) or confused_face
(π). The function performs a “fuzzy” search of the descriptions to find matching emojis.
The code above the highlighted lines extracts emoji descriptions from the event
object passed to the handler. The code below the highlighted lines wraps the result in a response with a proper status code for success or failure.
At the time of this writing, this sample Lambda function was deployed and available at the following HTTP endpoint:
curl -s -X POST --header "Content-type:application/json" \
--data '{"descriptions":["flame","confused"]}' \
https://64856ijzmi.execute-api.us-west-2.amazonaws.com/default/fuzz-emoji | \
json_pp
Invoking the function returns the following result:
{
"confused" : "('confused_face', 'π')",
"flame" : "('fire', 'π₯')"
}
The function successfully returns the fire
(π₯) emoji for the description “flame”, and the confused_face
emoji (π) for the description “confused.β
Isolating the AWS Lambda Code from AWS Specifics
To migrate the Lambda function to Knative, the core application logic must be decoupled from AWS-specific dependencies. In this case, the function’s main logic is already isolated. The get_emojis()
method only accepts a list of strings as input, which makes it more adaptable for other platforms.
If the get_emojis()
method were dependent on the AWS Lambda event
object, it would not be compatible with Knative and would require some refactoring, as Knative does not provide an event
object.
Migrating a Single-File Function to a Knative Function
The core logic of the function is encapsulated into a single Python module named fuzz_emoji.py
, which can be migrated to your Knative function.
Using a text editor of your choice, create the
fuzz_emoji.py
file in theget-emojis
directory:nano ~/get-emojis/fuzz_emoji.py
Give the file the following content:
- File: ~/get-emojis/fuzz_emoji/py
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
from typing import List, Mapping, Tuple import emoji import requests class FuzzEmoji: def __init__(self): self.emoji_dict = {} emoji_list = {name: data for name, data in emoji.EMOJI_DATA.items() if 'en' in data} for emoji_char, data in emoji_list.items(): name = data['en'].strip(':') self.emoji_dict[name.lower()] = emoji_char @staticmethod def get_synonyms(word): response = requests.get(f"https://api.datamuse.com/words?rel_syn={word}") if response.status_code == 200: synonyms = [word_data['word'] for word_data in response.json()] return synonyms raise RuntimeError(response.content) def get_emoji(self, description) -> Tuple[str, str]: description = description.lower() # direct match if description in self.emoji_dict: return description, self.emoji_dict[description] # Subset match for name in self.emoji_dict: if description in name: return name, self.emoji_dict[name] synonyms = self.get_synonyms(description) # Synonym match for syn in synonyms: if syn in self.emoji_dict: return syn, self.emoji_dict[syn] return '', '' def get_emojis(self, descriptions: List[str]) -> Mapping[str, str]: return {d: str(self.get_emoji(d)) for d in descriptions}
When complete, save your changes.
Run the
ls
command:ls -laGh ~/get-emojis/
The folder structure should now look like this:
total 52K drwxr-xr-x 3 USERNAME 4.0K Oct 10 17:32 . drwxr-x--- 9 USERNAME 4.0K Oct 10 16:51 .. -rwxr-xr-x 1 USERNAME 55 Oct 10 16:51 app.sh drwxrwxr-x 3 USERNAME 4.0K Oct 10 17:20 .func -rw-r--r-- 1 USERNAME 217 Oct 10 16:51 .funcignore -rw-r--r-- 1 USERNAME 1.8K Oct 10 16:51 func.py -rw-r--r-- 1 USERNAME 317 Oct 10 17:22 func.yaml -rw-rw-r-- 1 USERNAME 1.4K Oct 10 17:32 fuzz_emoji.py -rw-r--r-- 1 USERNAME 235 Oct 10 16:51 .gitignore -rw-r--r-- 1 USERNAME 28 Oct 10 16:51 Procfile -rw-r--r-- 1 USERNAME 862 Oct 10 16:51 README.md -rw-r--r-- 1 USERNAME 28 Oct 10 16:51 requirements.txt -rw-r--r-- 1 USERNAME 259 Oct 10 16:51 test_func.py
Edit your
func.py
file so that it calls thefuzz_emoji
module:nano ~/get-emojis/func.py
Insert or adjust the highlighted lines so that the contents of your
fuzz_emoji.py
file appear as below. Remember to save your changes:- File: ~/get-emojis/func.py
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67
from parliament import Context from flask import Request import json from fuzz_emoji import FuzzEmoji # parse request body, json data or URL query parameters def payload_print(req: Request) -> str: if req.method == "POST": if req.is_json: return json.dumps(req.json) + "\n" else: # MultiDict needs some iteration ret = "{" for key in req.form.keys(): ret += '"' + key + '": "'+ req.form[key] + '", ' return ret[:-2] + "}\n" if len(ret) > 2 else "{}" elif req.method == "GET": # MultiDict needs some iteration ret = "{" for key in req.args.keys(): ret += '"' + key + '": "' + req.args[key] + '", ' return ret[:-2] + "}\n" if len(ret) > 2 else "{}" # pretty print the request to stdout instantaneously def pretty_print(req: Request) -> str: ret = str(req.method) + ' ' + str(req.url) + ' ' + str(req.host) + '\n' for header, values in req.headers.items(): ret += " " + str(header) + ": " + values + '\n' if req.method == "POST": ret += "Request body:\n" ret += " " + payload_print(req) + '\n' elif req.method == "GET": ret += "URL Query String:\n" ret += " " + payload_print(req) + '\n' return ret def main(context: Context): """ Function template The context parameter contains the Flask request object and any CloudEvent received with the request. """ # Add your business logic here print("Received request") if 'request' in context.keys(): ret = pretty_print(context.request) print(ret, flush=True) descriptions = context.request.args.get('descriptions').split(',') fuzz_emoji = FuzzEmoji() result = fuzz_emoji.get_emojis(descriptions) return json.dumps(result, ensure_ascii=False), 200 else: print("Empty request", flush=True) return "{}", 200
Below is a breakdown of the file code functionality:
- Imports the built-in
json
, theContext
from parliament (the function invocation framework that Knative uses for Python functions), and theFuzzEmoji
class. - The
main()
function accepts the parliamentContext
as its only parameter, which contains a Flaskrequest
property. - The first line extracts the emoji descriptions from the Flask
request
arguments. It expects the descriptions to be a single comma-separated string, which it splits into a list ofdescriptions
. - Instantiates a
FuzzEmoji
object and calls theget_emojis()
method. - Uses the
json
module to serialize the response and return it with a200
status code.
Next, edit the
requirements.txt
file to include the dependencies offuzz_emoji.py
(therequests
andemoji
packages) in the Docker image:nano ~/get-emojis/requirements.txt
Append the highlighted lines to the end of the file, and save your changes:
- File: ~/get-emojis/requirements.txt
1 2 3
parliament-functions==0.1.0 emoji==2.12.1 requests==2.32.3
Re-build and re-deploy the container:
func build --registry docker.io/DOCKER_HUB_USERNAME func deploy
Test your function using the public URL:
curl http://get-emojis.default.IP_ADDRESS.sslip.io/?descriptions=cold,plane,fam
The
descriptions
provided as a query parameter are echoed back, along with a corresponding emoji name and emoji for each description:{"cold": "('cold_face', 'π₯Ά')", "plane": "('airplane', 'β')", "fam": "('family', 'πͺ')"}
This confirms that the Knative function works as expected.
Migrating a Multi-File Function to a Knative Function
In the previous example, the entire application logic was contained in a single file called fuzz_emoji.py
. For larger workloads, your function may involve multiple files or multiple directories and packages.
Migrating such a setup to Knative follows a similar process:
Copy all relevant files and directories into the same
get-emojis
directory.Import any required modules in
func.py
.Update the
requirements.txt
file to include all of the dependencies used across any of the modules.
Migrating External Dependencies
When migrating an AWS Lambda function, it may depend on various AWS services such as S3, DynamoDB, or SQS. It’s important to evaluate each dependency to determine the best option to suit your situation.
There are typically three options to consider:
Keep it as-is: Continue using the Knative function to interact with the AWS service.
Replace the service: For example, you might switch from an AWS service like DynamoDB to an alternative key-value store in the Kubernetes cluster.
Drop the functionality: Eliminate certain AWS-specific functionalities, such as no longer writing messages to AWS SQS.
Namespace and Service Account
The Knative function eventually runs as a pod in the Kubernetes cluster. This means it runs in a namespace and has a Kubernetes service account associated with it. These are determined when you run the func deploy
command. You can specify them using the -n
(or --namespace
) and --service-account
arguments.
If these options are not specified, the function deploys in the currently configured namespace and uses the default service account of the namespace.
If your Knative function needs to access any Kubernetes resources, itβs recommended to explicitly specify a dedicated namespace and create a dedicated service account. This is the preferred approach since it avoids granting excessive permissions to the default service account.
Configuration and Secrets
If your AWS Lambda function uses ParameterStore
and SecretsManager
for configuration and sensitive information, these details should not be embedded directly in the function’s image. For example, if your function needs to access AWS services, it would require AWS credentials to authenticate.
Kubernetes offers the ConfigMap
and Secret
resources for this purpose. The migration process involves the following steps:
Identify all the parameters and secrets the Lambda function uses.
Create corresponding
ConfigMap
andSecret
resources in the namespace for your Knative function.Grant the service account for your Knative function permissions to read
ConfigMap
andSecret
.
Roles and Permissions
Your Knative function may need to interact with various Kubernetes resources and services during migration, such as data stores, ConfigMaps
, and Secrets
. To enable this, create a dedicated role with the necessary permissions and bind it to the function’s service account.
If your architecture includes multiple Knative functions, it is considered a best practice to share the same service account, role, and role bindings between all the Knative functions.
Logging, Metrics, and Distributed Tracing
The logging experience in Knative is similar to printing something in your AWS Lambda function. In AWS Lambda, output is automatically logged to CloudWatch. In Knative, that same print statement automatically sends log messages to your container’s logs. If you have centralized logging, these messages are automatically recorded in your log system.
LKE provides the native Kubernetes dashboard by default. It runs on the control plane, so it doesn’t take resources from your workloads. You can use the dashboard to explore and monitor your entire cluster:
For production systems, consider using a centralized logging system like ELK/EFK, Loki, or Graylog, along with an observability solution consisting of Prometheus and Grafana. You can also supplement your observability by leveraging a telemetry data-oriented solution such as OpenTelemetry. These tools can enhance your ability to monitor, troubleshoot, and optimize application performance while ensuring reliability and scalability.
Knative also has built-in support for distributed tracing, which can be configured globally. This means your Knative function automatically participates in tracing without requiring additional changes.
The Debugging Experience
Knative offers debugging at multiple levels:
- Unit test your core logic
- Unit test your Knative function
- Invoke your function locally
When you create a Python Knative function, Knative generates a skeleton for a unit test called test_func.py
. At the time of this writing, the generated test is invalid and requires some modifications to work correctly. See this GitHub issue
for details.
Open the
test_func.py
file in theget-emojis
directory:nano ~/get-emojis/test_func.py
Replace its content with the test code below, and save your changes. This code is updated for testing the fuzzy emoji search functionality:
- File: ~/get-emojis/test_func.py
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
import unittest from parliament import Context func = __import__("func") class DummyRequest: def __init__(self, descriptions): self.descriptions = descriptions @property def args(self): return dict(descriptions=self.descriptions) @property def method(self): return 'GET' @property def url(self): return 'http://localhost/' @property def host(self): return 'localhost' @property def headers(self): return {'Content-Type': 'application/json'} class TestFunc(unittest.TestCase): # noinspection PyTypeChecker def test_func(self): result, code = func.main(Context(DummyRequest('flame,confused'))) expected = """{"flame": "('fire', 'π₯')", "confused": "('confused_face', 'π')"}""" self.assertEqual(expected, result) self.assertEqual(code, 200) if __name__ == "__main__": unittest.main()
Use
pip3
to install the dependencies listed in therequirements.txt
file:pip3 install -r ~/get-emojis/requirements.txt
Use the
python3
command to run thetest_func.py
file and test the invocation of your function:python3 ~/get-emojis/test_func.py
A successful test should produce the following output:
Received request GET http://localhost/ localhost Content-Type: application/json URL Query String: {"descriptions": "flame,confused"} . ---------------------------------------------------------------------- Ran 1 test in 0.395s OK
Once the code behaves as expected, you can test the function locally by packaging it in a Docker container using func invoke
to run it. This approach is handled completely through Docker, without the need for a local Kubernetes cluster.
After local testing, you may want to optimize the function’s image size by removing any redundant dependencies to improve resource utilization. Deploy your function to a staging environment (a Kubernetes cluster with Knative installed) using func deploy
. In the staging environment, you can conduct integration, regression, and stress testing.
See More Information below for resources to help you get started with migrating AWS Lambda functions to Knative functions on the Linode Kubernetes Engine (LKE).
More Information
You may wish to consult the following resources for additional information on this topic. While these are provided in the hope that they will be useful, please note that we cannot vouch for the accuracy or timeliness of externally hosted materials.
- Knative
- Knative Functions
- Knative Functions - Deep Dive (Video)
- Accessing request traces - Knative
- Migrating from AWS Lambda to Knative Functions
- GitHub: boson-project/parliament
- Logging and Metrics with Amazon CloudWatch
- Prometheus
- Grafana Labs - Loki, Grafana, Tempo, Mimir
- OpenTelemetry
- Sample AWS Lambda function
- Sample Knative function (Python)
This page was originally published on