The Journey to Mastering Continuous Delivery:

The Challenges

• Manual processes that involve deploying and managing infrastructure from local developer machines.
• Incidents in production caused by untracked deployments.

The Goals

• Reevaluate delivery platforms and solutions following a Kubernetes migration. Avoid building something in-house would have taken away from engineering resources to work on business/product differentiators and features.
• Find a solution with APM and ELK stack integration which is integral for verifying services in production and tracking incident data.
• Avoid having to track incident data on local machines with Fabric deployments manually.

The Dream

• Integrate with the ELK stack architecture, mapping events to deployments. “This allows engineers to compare the baseline error rate with any new error rates created by an application deployment.”
• Audit trails for easy change management tracking.
• Standardize and automated deployments.
• Reduce the number of change incidences.

Advice

• Having an incident management tool is paramount to succeeding in the SRE role.
• Without a management system, teams will struggle to learn from their mistakes.
• Tenzin recommends “implementing an error budget system, where each team has an error threshold before they must pause and focus solely on fixing existing problems.”

“The Harness platform standardizes the way Iterable deploys, so deployments are no longer a mental drain… Since Harness, we’ve reduced the number of change incidences by 20%.”

About the Heroes Program

To celebrate the leaders, the visionaries, and the contributors, The Heroes of CD program shares the accomplishments of industry professionals…

Read the blog >