Non-adherence to DevOps practices would be challenging for an organization, for the following reasons:
- High deployment effort for each of the development, QA, and production systems
- Complex manual installation procedures are cumbersome and expensive
- Lack of a comprehensive operations manual makes the system difficult to operate
- Insufficient trace or log file details makes troubleshooting incomplete
- Application-specific issues of performance impact not assessed for other applications
- SLA adherence, as required by the business application, would be challenging
- Monitoring servers, filesystems, databases, and applications in isolation will have gaps
- Business application redundancy for failover is expensive in isolation
DevOps adoption and maturity for big data systems will benefit organizations in the following ways:
- DevOps processes can be implemented as standalone or a combination of other processes
- Automation frameworks will improve business efficiency
- DevOps frameworks will help to build resilience into the application's code
- DevOps processes incorporate SLAs for operational requirements
- The operations manual (runbook) is prepared in development to aid operations
- In matured DevOps processes, runbook-driven development is integrated
- In DevOps processes, application-specific monitoring is part of the development process
- DevOps planning considers high availability and disaster recovery technology
- Resilience is built into the application code in-line with technology features
- DevOps full-scripted installation facilitates fully automate deployment
- DevOps operation team and developers are familiar with using logging frameworks
- The non-functional requirements of operability, maintenance, and monitoring get sufficient attention, along with system development specifications
- Continuous integration and continuous delivery eliminates human errors, reduces planned downtime for upgrades, and facilitates productivity improvements