Import Infrastructure Objects via AWS

Infrastructure Catalog allows you to import Infrastructure Objects across your various AWS accounts. Once imported, you can see details about your objects (ARN, source account, region, storage size, etc), filter across AWS accounts ("how many databases do I have cumulatively across all my accounts?"), and define relationships between infrastructure objects to Services, Systems, and other Infrastructure Objects.

Integrating an AWS account with OpsLevel

To integrate your AWS account with OpsLevel, you must first create an IAM policy and IAM role in your AWS account. Once created you must configure the role with an AWS External ID generated by OpsLevel.

Setup

Generate an external ID

To generate an external ID, navigate to Add New Integration from Integrations in the side navigation. Once on the New Integration page, find and select AWS. A modal will appear where you can generate and copy a UUID. Keep note of this ID for later.

Create an AWS IAM policy

After generating the external ID, you'll need to create an IAM policy for the OpsLevel role in your AWS account with the necessary permissions:

1. Create a new policy in the AWS IAM Console
2. Select the JSON tab. Paste the following permission policies in the textbox:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "autoscaling:Describe*",
                "cloudfront:GetDistribution",
                "cloudfront:ListDistributions",
                "dynamodb:Describe*",
                "dynamodb:List*",
                "ec2:Describe*",
                "ec2:Get*",
                "ecs:Describe*",
                "ecs:List*",
                "elasticloadbalancing:Describe*",
                "elasticache:Describe*",
                "elasticache:List*",
                "es:Describe*",
                "es:List*",
                "es:Get*",
                "grafana:Describe*",
                "grafana:List*",
                "kafka:Describe*",
                "kafka:List*",
                "kinesis:Get*",
                "kinesis:List*",
                "kinesis:Describe*",
                "lambda:Get*",
                "lambda:List*",
                "rds:Describe*",
                "rds:List*",
                "redshift:Describe*",
                "redshift:List*",
                "route53domains:Get*",
                "route53domains:List*",
                "s3:Describe*",
                "s3:List*",
                "s3:GetBucketLocation",
                "s3:GetBucketTagging",
              	"s3:GetBucketPolicyStatus",
                "s3:GetBucketVersioning",
                "sns:Get*",
                "sns:List*",
                "storagegateway:List*",
                "storagegateway:Describe*",
                "sqs:Get*",
                "sqs:List*",
                "tag:Get*",
                "waf:Get*",
                "waf:List*",
                "wafv2:Get*",
                "wafv2:List*",
                "wafv2:Describe*"
            ],
            "Resource": [
                "*"
            ]
        }
    ]
}

3. Click Next: Tags and Next: Review
4. Name the policy OpsLevelIntegrationPolicy or one of your own choosing. Provide a description.
5. Click Create policy

Create an AWS IAM role

Create an IAM role for OpsLevel to use the permissions defined in the IAM policy.

1. Create a new role in the AWS IAM Console.
2. Select AWS account for the trusted entity type, and Another AWS account.
3. Enter 746108190720 as the Account ID. This is OpsLevel’s account ID, and grants OpsLevel access to your AWS data.
4. Select Require external ID and enter the external ID generated in the previous step.
5. Click Next.
6. If you’ve already created the policy, search for it on this page and select it. Otherwise, click Create Policy, which opens in a new window, and follow the instructions from the previous section.
7. Click Next.
8. Give the role a name such as OpsLevelIntegrationRole, as well as an apt description.
9. Click Create Role.

Sample Trust Relationship JSON:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::746108190720:user/opslevel-integration"
            },
            "Condition": {
                "StringEquals": {
                    "sts:ExternalId": "XXXXXXXXXXXXXX"
                }
            },
            "Action": "sts:AssumeRole"
        }
    ]
}

Install via GraphQL

Once your IAM Role and Policy are created. You can create the integration in GraphQL.

Verify AWS integration was successful

Once you have integrated your AWS account(s) you should now be able see them by navigating to Installed Integrations via Integrations on the side navigation. Once you've clicked on your integration you will be able to see details about your integration such as the Account ID and when it was last synced.

Inferring ownership via AWS tags

Many AWS customers utilize tags within the admin console to assign ownership to Infrastructure Objects. Our AWS Integration allows you to leverage these existing tags to make ownership assignment for Infrastructure Objects simple. By default, when a new integration is created we set a default ownership defining tag key of owner. If the value of owner matches a Team or Group alias we will automatically assign ownership of that Infrastructure Object to that Team or Group. Ownership for these objects will be locked from the UI or API.

You can use GraphQL to update the array of tag keys. You can also use GraphQL to disable inferred ownership if you wish to assign it manually instead using the awsTagsOverrideOwnership field.

Here's a sample mutation updating the ownership inference via tags:

 mutation  {
  awsIntegrationUpdate(integration: {id: "XXX"}, input: {ownershipTagKeys: ["owner","service","app"], awsTagsOverrideOwnership: true}) {
    integration {
      id
      iamRole
      externalId
      aliases
      ownershipTagKeys
      awsTagsOverrideOwnership
    }
    errors {
      message
      path
    }
  }
}

What Infrastructure objects and types are imported?

We support a wide range of commonly used AWS Infrastructure Objects and Types as part of our integration, namely:

• Cache
  • ElastiCache
• Compute
  • EC2
• Database
  • DynamoDB
  • RDS
• Networking
  • VPC
• Queues / Message Brokers
  • SNS
  • SQS
• Serverless Function
  • Lambda
• Storage
  • S3

If there is an Object or Type not listed that you would be interested in importing, please reach out to our customer support team.

We synchronize changes from AWS every 12 hours.

Accessing Infrastructure objects

You can access your imported infrastructure objects by selecting Infrastructure in Catalog from the side navigation.

Once in the Infrastructure list page you will see a table listing all of the infrastructure objects you have imported. At a glance you will be able to see the following data around your objects:

• Name
• Resource Type
• Zone
• Source
• Tags
• Owner

You can discover even more information about your Infrastructure Objects by clicking the View Details icon in the table. This will open a drawer component that will display when the object was last synced along with a host of attributes we've pulled in from AWS around that object.

For a full breakdown of what we've synced from your AWS account, you can select the Raw JSON tab.

Selecting the name of the object from the list will open a new details page for this object. The Infrastructure Details page contains all of the information displayed in the View Details drawer with the added bonus that you can also see which Services and other Infrastructure your infrastructure object relates to. See "Adding Relationships to Infrastructure Objects" further down in the docs.

Filtering on Infrastructure objects

Similar to Services, you can apply filters against your list of Infrastructure Objects. Click the Filter icon from the Infrastructure List page to open filters. Filters work the same for Infrastructure as they do elsewhere in OpsLevel with the added bonus that you can filter on the attributes of your infrastructure objects. As such, you can filter for queries like "which of my databases have multi-zone availability enabled?".

Adding relationships to Infrastructure objects

Utilizing both GraphQL and AWS tags, you can add relationships to your Infrastructure Objects in three ways:

• "depends on"
  • Infrastructure Objects can "depend on" Services or other Infrastructure Objects. A "depends on" relationship implies that the Infrastructure Object relies on the given Service or other Infrastructure Object to operate.
• "dependency of"
  • Infrastructure Objects can be "dependent of" Services or other Infrastructure Objects. A "dependent of" relationship implies that the the Service(s) or other Infrastructure Object(s) related to your chosen Infrastructure Object rely on it to operate.
• "belongs to"
  • Infrastructure Objects can belong to a System. A "belongs to" relationship implies that your chosen Infrastructure Object is part of a System that forms a single unit.

Establishing relationships between Infrastructure objects and Services/Other Infrastructure objects

Using GraphQL, you can establish that a given Infrastructure Object can depend on or be the dependent of either another Infrastructure Object(s) and/or Service(s). See the following example in GraphQL:

mutation relationshipCreate($source: IdentifierInput!, $target: IdentifierInput!, $type: RelationshipTypeEnum!) {
        relationshipCreate(relationshipDefinition: {
          source: $source,
          target: $target,
          type: $type
        }) {
          relationship {
            id
            type
            source {
              ... on Service {
                id
                name
                 __typename
              }
              ... on System {
                id
                name
                 __typename
              }
              ... on InfrastructureResource {
                id
                name
                __typename
              }
            }
            target {
              ... on Service {
                id
                name
                 __typename
              }
              ... on System {
                id
                name
                 __typename
              }
              ... on Domain {
                id
                name
                 __typename
              }
              ... on InfrastructureResource {
                id
                name
                 __typename
              }
            }
          }
          errors {
            message
            path
          }
        }
      }

Once you have set the relationship for your Infrastructure Object, you can see it in two places. The first is within the "Relationship" table in the Infrastructure Details page.

The second can be found by navigating to the Services show page and clicking on the Infrastructure tab under Dependencies

Establishing relationships between Infrastructure objects and Systems

Using GraphQL, you can establish which System a given Infrastructure Object belongs to. See the following example in GraphQL:

mutation relationshipCreate($source: IdentifierInput!, $target: IdentifierInput!, $type: RelationshipTypeEnum!) {
        relationshipCreate(relationshipDefinition: {
          source: $source,
          target: $target,
          type: $type
        }) {
          relationship {
            id
            type
            source {
              ... on Service {
                id
                name
                 __typename
              }
              ... on System {
                id
                name
                 __typename
              }
              ... on InfrastructureResource {
                id
                name
                __typename
              }
            }
            target {
              ... on Service {
                id
                name
                 __typename
              }
              ... on System {
                id
                name
                 __typename
              }
              ... on Domain {
                id
                name
                 __typename
              }
              ... on InfrastructureResource {
                id
                name
                 __typename
              }
            }
          }
          errors {
            message
            path
          }
        }
      }

Once you have set the relationship between your Infrastructure Object and System, you will be able to see that your Infrastructure Object now belongs to the System by navigating to the Systems Display page and clicking on the Infrastructure tab under Relationships

Establishing relationships for Infrastructure Objects via AWS tags

Customers can utilize tags within the admin console to assign relationships to Infrastructure Objects. Our AWS Integration allows you to leverage these tags to make relationship assignment for Infrastructure Objects simple. We allow customers to set three different tags:

• opslevel-depends-on
• opslevel-dependency-of
• opslevel-belongs-to

For opslevel-depends-on and opslevel-dependency-of, if the value matches a Service or Infrastructure Object alias, we will automatically assign that relationship. We automatically generate an alias for AWS Objects that is their ARN so that they can be easily used for the tags. We will also lock the ability to adjust these relationships from the API.

Note: Only one value can be assigned per key. If you wish to assign multiple relationships you can add multiple of the same key with unique suffixes appended in the format of opslevel-depends-on* or opslevel-dependency-of* (e.g. opslevel-depends-on1:shopping_cart, opslevel-depends-on2:catalog_service). Keep in mind that AWS has a 50 tag per object limit.

For opslevel-belongs-to, if the value matches a System alias, we will automatically assign that relationship.

Note: Infrastructure Objects can only belong to a single System. As such you are restricted from utilizing multiple opslevel-belongs-to tags. These tags will also take precedent over relationships defined via the API.

Using the OpsLevel CLI to manage your AWS integration

In addition to the GraphQL API, you can also manage your AWS Integration with the following options:

• Terraform provider
• Command Line Utility

Example: Create a new AWS integration via CLI

cat << EOF | opslevel create integration -f -
kind: aws
spec:
  name: "Prod"
  iamRole: "arn:aws:iam::XXXXX:role/opslevel-integration"
  externalId: "XXXXXX"
  ownershipTagOverrides: true
  ownershipTagKeys: ["owner","service","app"]
EOF

Example: Create a new AWS integration via Terraform

resource "opslevel_integration_aws" "dev" {
  name = "dev"
  iam_role = aws_iam_role.opslevel.arn
  external_id = random_string.external_id.result
  ownership_tag_overrides = true
  ownership_tag_keys = ["owner","team","unit"]
}

Example: Infer Ownership through AWS tags via CLI

cat << EOF | opslevel update integration XXXXXXXX -f -
kind: aws
spec:
  ownershipTagOverrides: true
  ownershipTagKeys: ["owner","service","app"]
EOF