jessflan / cfn-datadog

🐕⚙️Cloudformation custom resources that integrate with datadog

README

What is this repository for?

• To setup datadog monitoring via cloudformation

What does it do?

It setups a lambda function that communicate with the monitors and timeboard resource on the datadog api

https://docs.datadoghq.com/api/#monitors https://docs.datadoghq.com/api/#timeboards

The lambda is invoked via cloudformation custom resource. This repo supplies custom resource classes for all the 4 Monitor types. Custom resources have the same exact data structure as the DD api

It uses https://github.com/nordcloud/cfn-encrypt to securely store datadog api credentials

This repository makes use of Troposphere

How do I get set up?

• clone this repo

git clone [email protected]:nordcloud/cfn-datadog.git

• Build the lambda package

chmod +x create_zip.sh && ./create_zip.sh

• upload datadog_lambda.zip to an s3 bucket
• get your api and application keys from Datadog
• generate the lambda template

python lambda-template.py > /tmp/lambda.template

• If you have not already set up your encrypt stack do that now ( https://github.com/nordcloud/cfn-encrypt )
• create a stack from your lambda.template file
• The lambda is now set up and ready to use.

How do i use it to monitor resources created in other stacks?

Install the cfn_datadog library

pip install cfn-encrypt

Import the custom resource classes you want to use

from cfn_datadog import (
    MetricAlert, MetricAlertOptions, Composite, CompositeOptions,
    EventAlert, EvenAlertOptions, ServiceCheck, ServiceCheckOptions,
    Timeboard
)

Add a parameter so that you can reference your lambda stack

datadog_lambda_stackname = t.add_parameter(Parameter(
    "DatadogLambdaStackname",
    Type="String",
    Description="Datadog lambda stackname",
))

Add the custom resource to the template: For documentation see datadog api

Example MetricAlert

t.add_resource(MetricAlert(
    'Alert2',
    query=Join("",["avg(last_1h):avg:system.net.bytes_rcvd{host:",Ref(my_instance),"} < 40"]),
    ServiceToken=ImportValue(
    Sub("${DatadogLambdaStackname}-LambdaArn")),
    name="Bytes received on mytesthost",
    message="Some Message @MyDDHandle",
    tags=["tag1", "tag2"],
    options=MetricAlertOptions(
        notify_no_data= True,
        no_data_timeframe=50
    )
))

Example Timeboard

t.add_resource(Timeboard(
    'ExampleTimeBoard',
    ServiceToken=ImportValue(Sub("${DatadogLambdaStackname}-TimeboardLambdaArn")),
    TimeboardTitle="testboard",
    description="Sample testboard",
    graphs=[Graph(
        definition=Definition(
            events=[],
            requests=[Request(
                q="avg:system.mem.free{*}"
            )],
            viz="timeseries"
        ),
        title="Average Memory Free"
    )],
    template_variables=[TemplateVariable(
        name="host1",
        prefix="host",
        default="host:my-host"
    )],
    read_only=True
))