Scenario
A large retail provider, has recently undergone a major digital transformation initiative, with Azure Kubernetes Service providing the primary hosting for customer facing services. Forming part of a broader digital shift within the business, Arinco was engaged to deliver a solution that could monitor and maintain the platform and to provide a long-term managed service.
Solution
Arinco leveraged its Azure Kubernetes Service (AKS) knowledge to assist the customer with implementing a modern, metrics-based monitoring solution for deployments hosted on AKS.
As part of this initiative, Arinco completed a review of the customer application architecture and advised on improvements. From there we completed the design, developed metrics, customer SLAs, and deployed a production ready monitoring solution onto Azure Kubernetes Service quickly and securely.
Key activities included:
- Further development of custom applications on the customer AKS platform.
- Custom builds of docker containers, with storage of artifacts located in Azure Container Registry.
- Designing, planning and implementation of a multi-environment AKS monitoring solution.
- Integration of Prometheus, capturing cluster level alerting
- Integration with Azure Monitor, using standard Alert Rules and Action Groups
- Integration of modern security products, such as Azure key vault to the customer deployment process.
- Implementation of both platform and application scaling using AKS cluster auto-scaling and pod auto-scaling.
- Developing infrastructure CI/CD pipelines for the AKS platform
- Developing application CI/CD pipelines for container-based workloads.
- Creating documentation of both the AKS and application solution, including general operations runbooks.
Outcomes
Over the past two years, Arinco has ensured that services are available according to customer defined SLA, with 24/7 alerting, monitoring and operations the production AKS environment. Arinco provided the following innovative outcomes:
- From week one, AKS Clusters were monitored for uptime, performance and issues, aligned to Microsoft and industry best practices.
- Modern DevOps standards were utilised within the customer environment and development process.
- Automatic deployment to each environment can be completed in less than 5 minutes.
- Issues are automatically triaged and sent to the appropriate stakeholder.
