Component: datadog-logs-archive

This component is responsible for provisioning Datadog Log Archives. It creates a single log archive pipeline for each AWS account. If the catchall flag is set, it creates a catchall archive within the same S3 bucket.

Each log archive filters for the tag env:$env where $env is the environment/account name (ie sbx, prd, tools, etc), as well as any tags identified in the additional_tags key. The catchall archive, as the name implies, filters for '*'.

A second bucket is created for cloudtrail, and a cloudtrail is configured to monitor the log archive bucket and log activity to the cloudtrail bucket. To forward these cloudtrail logs to datadog, the cloudtrail bucket's id must be added to the s3_buckets key for our datadog-lambda-forwarder component.

Both buckets support object lock, with overrideable defaults of COMPLIANCE mode with a duration of 7 days.

Prerequisites

• Datadog integration set up in target environment
  • We rely on the datadog api and app keys added by our datadog integration component

Issues, Gotchas, Good-to-Knows

Destroy/reprovision process

Because of the protections for S3 buckets, if we want to destroy/replace our bucket, we need to do so in two passes or destroy the bucket manually and then use terraform to clean up the rest. If reprovisioning a recently provisioned bucket, the two-pass process works well. If the bucket has a full day or more of logs, though, deleting it manually first will avoid terraform timeouts, and then the terraform process can be used to clean up everything else.

Two step process to destroy via terraform

• first set s3_force_destroy var to true and apply
• next set enabled to false and apply or use tf destroy

Usage

Stack Level: Global

Here's an example snippet for how to use this component. It's suggested to apply this component to all accounts from which Datadog receives logs.

components:
  terraform:
    datadog-logs-archive:
      settings:
        spacelift:
          workspace_enabled: true
      vars:
        enabled: true
 #       additional_query_tags:
 #         - "forwardername:*-dev-datadog-lambda-forwarder-logs"
 #         - "account:123456789012"

Requirements

Name	Version
terraform	>= 0.13.0
aws	>= 2.0
datadog	>= 3.3.0
local	>= 1.3

Providers

Name	Version
aws	>= 2.0
datadog	>= 3.7.0
http	>= 2.1.0

Modules

Name	Source	Version
cloudtrail	cloudposse/cloudtrail/aws	0.21.0
cloudtrail_s3_bucket	cloudposse/cloudtrail-s3-bucket/aws	0.23.1
iam_roles	../account-map/modules/iam-roles	n/a
s3_bucket	cloudposse/s3-bucket/aws	0.46.0
this	cloudposse/label/null	0.25.0

Resources

Name	Type
aws_caller_identity.current	data source
aws_partition.current	data source
aws_ssm_parameter.datadog_api_key	data source
aws_ssm_parameter.datadog_app_key	data source
aws_ssm_parameter.datadog_aws_role_name	data source
aws_ssm_parameter.datadog_external_id	data source
datadog_logs_archive.catchall_archive	resource
datadog_logs_archive.logs_archive	resource
http.current_order	data source

Inputs

Name	Description	Type	Default	Required
additional_query_tags	Additional tags to include in query for logs for this archive	list	[]	no
catchall	Set to true to enable a catchall for logs unmatched by any queries. This should only be used in one environment/account	bool	false	no
datadog_aws_account_id	The AWS account ID Datadog's integration servers use for all integrations	string	464622532012	no
enable_glacier_transition	Enable/disable transition to glacier. Has no effect unless lifecycle_rules_enabled set to true	bool	true	no
glacier_transition_days	Number of days after which to transition objects to glacier storage	number	365	no
lifecycle_rules_enabled	Enable/disable lifecycle management rules for s3 objects	bool	true	no
object_lock_days_archive	Set duration of archive bucket object lock	number	7	yes
object_lock_days_cloudtrail	Set duration of cloudtrail bucket object lock	number	7	yes
object_lock_mode_archive	Set mode of archive bucket object lock	string	COMPLIANCE	yes
object_lock_mode_cloudtrail	Set mode of cloudtrail bucket object lock	string	COMPLIANCE	yes
s3_force_destroy	Set to true to delete non-empty buckets when enabled is set to false	bool	false	for destroy only

Outputs

Name	Description
archive_id	The ID of the environment-specific log archive
bucket_arn	The ARN of the bucket used for log archive storage
bucket_domain_name	The FQDN of the bucket used for log archive storage
bucket_id	The ID (name) of the bucket used for log archive storage
bucket_region	The region of the bucket used for log archive storage
cloudtrail_bucket_arn	The ARN of the bucket used for cloudtrail log storage
cloudtrail_bucket_domain_name	The FQDN of the bucket used for cloudtrail log storage
cloudtrail_bucket_id	The ID (name) of the bucket used for cloudtrail log storage
catchall_id	The ID of the catchall log archive

References

• cloudposse/s3-bucket/aws - Cloud Posse's S3 component
• [datadog_logs_archive resource] (https://registry.terraform.io/providers/DataDog/datadog/latest/docs/resources/logs_archive) - Datadog's provider documentation for the datadog_logs_archive resource