Running many services.
WHAT CHANGES
Deployment becomes many deployments Configuration multiplies Failure modes multiply
WHAT TO AUTOMATE
Everything.
WHY
Manual processes do not scale to many services.
WHAT TO STANDARDISE
Build and deployment process Logging format Health endpoints Configuration approach Metrics
WHY
So operating one service teaches you to operate all of them.
WHAT CONTAINERS PROVIDE
Consistent packaging and environment.
WHAT ORCHESTRATION PROVIDES
Scheduling, scaling, restarting and routing.
WHAT IT COSTS
Substantial operational complexity.
WHAT TO ESTABLISH
Whether you have the capacity to operate it.
WHY
Orchestration platforms require ongoing expertise, and are frequently adopted without it.
WHAT TO IMPLEMENT
Health checks distinguishing started from ready Graceful shutdown Resource limits per service
WHAT TO PLAN
How a service is rolled back, quickly.
WHAT TO TEST
Failure: what happens when a dependency is unavailable.
