DevOps – 3 factors for success

ONE team - Working together as a team and with the customer

The “ONE team” concept has two sides:

• No specialization, but building up of broad knowledge in the team
• No sub-teams that deal exclusively with individual sub-areas; the entire team works together on all tasks that arise

Building broad knowledge is a prerequisite to enable all team members to meet team responsibilities. The opportunity for personal development is significantly increased for everyone.

All tasks, but especially those from the Ops part, are taken on as rotating responsibilities. The combination of development and operational knowledge has decisive advantages:

• There are no frictional losses in communication between operational and development responsibilities.
• The professional and technical expertise of development enables faster analysis and correction of errors.
• The direct feedback of knowledge from operational tasks as improvement of code, development process and test procedure is easily possible.

An important pillar in the sustainable development of such teams is a regular knowledge sharing and the transfer of expert knowledge. In practice, technical/professional presentations and workshops in a cycle of up to three weeks have proven effective.

Open, transparent and trusting communication contributes enormously to success. We have had very positive experiences, for example, with short, daily "health meetings" in which the results of application monitoring and current anomalies are discussed. All stakeholders (product owner, business consultants, architect, developer, scrum master) are invited. The direct exchange significantly speeds up the analysis of problems. The team identifies very strongly with the product and the sense of shared responsibility is reinforced.

Appropriate tools – Using tools for development, infrastructure and monitoring correctly

Tools should be selected specifically according to application purpose and use in order to best support the technologies and infrastructures used. This also includes the targeted selection of features and configurations within individual tools.

The following is an excerpt of tools from practical experience: 

• Communication & Specification - Jira, Confluence, Microsoft Teams etc.
  • These are particularly important for incident management and short coordination paths.
• Development & Infrastructure – Docker Desktop, Lens, Infrastructure-As-Code (Helm, Terraform etc.)
  • The choice of tools may vary depending on the cloud provider and the technology of the application.
• Build management - Jenkins, Azure DevOps, Github Actions etc.
  • For a high degree of automation of the CI/CD process and the development of the infrastructure, the appropriate tools must be selected.
• Monitoring – Grafana, Kibana, Prometheus, Jaeger, Opentelemetry etc.
  • Meaningful monitoring of the application includes access to the productive environment, for example, to check logs, dashboards and alerts, as well as for early error detection.
• Security & FOSS – Blackduck, Sonar, static code analysis, etc.
• It is important not to neglect the security and handling of open source libraries and to perform automated checks accordingly.