How Intuit Monitors Connectivity to AWS
0 likes2,554,416 views
The document discusses how Intuit monitors connectivity to AWS for its Turbotax services, detailing the evolution from monolithic to microservice architecture. It highlights the challenges of monitoring in a hybrid cloud environment and the use of ThousandEyes for synthetic monitoring to track connectivity issues without relying solely on application performance. Future steps include expanding monitoring capabilities to other infrastructure services within AWS.
How Intuit Monitors Connectivity to AWS
- 1. Dan Shanahan Staff Network Performance Engineer – TurboTax.com June 8th 2017 How Intuit Monitors Connectivity to AWS
- 2. 2
- 3. 3 AWS Hosted Data Centers • Consolidated and fewer data centers • Cloud-based services • Microservice architecture • Continuous integration/Continuous deployment Intuit Hosted Data Centers • Multiple data centers • Monolithic code base • Manual deployment and configuration Intuit Hosting Journey A Few Years Ago Today
- 4. 4 Intuit Hosting Journey TurboTax.com Utilization Patterns
- 5. 5 Intuit Application Architecture Service Oriented Architecture TurboTax WebTurboTax Web Tax Return Filing Customer Data Storage Tax Return Filing Customer Data Storage
- 6. How do we monitor these services?
- 8. 8 Monitoring Hybrid Cloud Why is it difficult? TurboTax WebTurboTax Web Tax Return Filing Customer Data Storage Tax Return Filing Customer Data Storage
- 9. 9 The Canary in the Coal Mine
- 10. 10 ThousandEyes Deployment AWS Synthetic Monitoring • EC2 Instances running Apache presenting a 1MB random data file • S3 presenting 1MB random data file • Deployed in every Availability Zone in all regions • Minimal services enabled to reduce false-positives – No DNS, load balancing, or downstream service reliance We’re monitoring the connectivity between AWS and Intuit.
- 12. 12 ThousandEyes Deployment Why not just monitor the service? •The app is architected to be fail-safe • Even if there is a connectivity problem, we may not see it because of the active/active nature of the applications •We might not know where the failure occurred because of multi-AZ deployments •Red herrings • Down-stream service failure • Traffic black-hole • ELB scaling events
- 14. 14 Visualizing ThousandEyes Data ThousandEyes Metrics ThousandEyes Metrics Time-series Metrics Wavefront/ThousandEyes Integration
- 16. 16 Visualizing ThousandEyes Data Wavefront Dashboards
- 17. 17 Visualizing ThousandEyes Data Wavefront Dashboards
- 18. 18 Visualizing ThousandEyes Data Wavefront Dashboards
- 19. 19 Visualizing ThousandEyes Data Wavefront Dashboards
- 20. 20 Next Steps Enterprise Agents in AWS Expand synthetic monitoring to other infrastructure services • DNS • NTP • Encryption services
- 21. Thank You!