Architecture Backcasting (EN)

Transcript

1. Success Criteria Risks Actions Day 1 Operation Day 2 Operation Test Release Dark Launch Monitoring Incident Response Success Criteria Development Design Develop Reduce Operational Cost Reduce development costs Reduce number of incidents by 20% Reduce error in nightly batch Clean code Lack of capacity (DB) Lack of capacity (# of concurrent connections) Target: x3 peak day container orchestration 12 factor apps (automation friendly) Decompose DB Nightly batch doesn't end in the expected timebox Reduce MTTR by 30% Reduce nightly batch with file integration and move on to realtime processing Can't identify the root cause distributed tracing Too many bugs caused by complexity Too many test cases Common utilities are very complex Must ask someone for details automation refactoring common utilities Takes longer time to fix an issue Shift Left Refine our testing strategy Lack of capacity planning process Capacity planning process/fra mework Get metrics for capacity planning (Observability) Tidy code Reduce human efforts by automation (90% of tasks) Different operational manuals required Different operational manuals required standardize operatonal tasks (it is easier to automate) Parallel processing (if we still need it)