Skip to content

Latest commit

 

History

History
75 lines (59 loc) · 3.17 KB

README.md

File metadata and controls

75 lines (59 loc) · 3.17 KB

AutoOps

This project contains useful operational processes represented as state machines with AWS StepFunctions.

flowchart TD
    CloudWatch["CloudWatch Alarm"] --> OpsItem;
    OpsItem --> |Manually start|Incident;
    Incident --> SNS;
	SNS --> ChatBot;
	SNS --> Lambda["Notification Lambda"];
	ChatBot --> Chime/Slack/Teams;
	Lambda --> Wechat/Wecom/Feishu/Dingding;
	Incident --> EventBridge;
	EventBridge --> StepFunction["AutoOps workflow"];
Loading
flowchart TD
    CloudWatch["CloudWatch Alarm"] --> Incident;
    Incident --> |Auto create|OpsItem;
    Incident --> SNS;
	SNS --> ChatBot;
	SNS --> Lambda["Notification Lambda"];
	ChatBot --> Chime/Slack/Teams;
	Lambda --> Wechat/Wecom/Feishu/Dingding;
	Incident --> EventBridge;
	EventBridge --> StepFunction["AutoOps workflow"];
Loading

How to deploy

You can pickup one region to deploy SNS topic

  1. Start CloudShell and clone this project

    ![https://console.aws.amazon.com/cloudshell/home]

    # cd ~
    # git clone https://github.com/lprincewhn/AutoOps.git
    
  2. Pick up a main region to deploy a common SNS topic where notifications will be sent to, or you can use an SNS topic existed by setting it in environment variable $SNS_TOPIC_ARN.

    # MAIN_REGION=<main region>
    # ACCOUNT=$(aws sts get-caller-identity --query 'Account' --output text)
    # cd ~/AutoOps/Common/
    # sam build
    # sam deploy --stack-name AutoOpsCommon --region $MAIN_REGION --confirm-changeset --resolve-s3 --capabilities CAPABILITY_IAM
    # SNS_TOPIC_ARN="arn:aws:sns:${MAIN_REGION}:${ACCOUNT}:AutoOps"
    # aws sns subscribe --topic-arn ${SNS_TOPIC_ARN}" --protocol email --notification-endpoint <[email protected]> 
    
  3. Go into sub-directory of each process to continue.

  • AlarmProcessor: Process CloudWatch alarms and send notifications
  • PhdEventProcessor: Process PHD events and send notifications
  • ComplianceProcessor: Process non-compliant resource in AWS Config and send notifications
  • EC2Provision: Create/Delete alarms for EC2 instances when they are started/terminated
  • RDSProvision: Create/Delete alarms for RDS database nodes when they are started/terminated
  • ESProvision: Deploy CloudWatch alarms for AWS Elasticsearch domain
  • CloudFrontProvision: Deploy CloudWatch alarms and update tags for AWS CloudFront distribution
  • RedshiftAuditLogToES: Transform and deliver Redshift audit logs in S3 bucket to Amazon ElasticSearch service
  • VisualizeCloudFrontLog: Transform and deliver CloudFront standard logs in S3 bucket to CloudWatch or OpenSearch service
  • TagCompliance: Check tags of EC2 periodly and notify users when tags are not compliant
  • MergeMetricData: Merge CloudWatch metric data to to centralized S3 bucket
  • SecurityHarden: Harden security by AWS Guardduty

How to try

The state machines, who represent operational processes, should be triggered by CloudWatch events (for AWS resource). You can use awscli command to start state machines' execution mannully.