Managing communication between services.
WHAT ONE IS
Infrastructure handling service-to-service communication, usually through proxies alongside each workload.
WHAT IT PROVIDES
Encryption between services, automatically Retries, timeouts and circuit breaking Traffic splitting for gradual rollouts Detailed metrics and tracing Policy about which services may call which
WHY THAT APPEALS
Those concerns are handled once rather than in every application.
WHAT IT COSTS
A proxy per workload, consuming resources Latency on every call Substantial complexity Another component to upgrade and diagnose
WHAT PROBLEMS IT ACTUALLY SOLVES
Many services, in several languages, needing consistent behaviour.
WHEN IT IS NOT JUSTIFIED
A handful of services One language, where a shared library would do A team without capacity to operate it
WHAT TO DO INSTEAD AT SMALL SCALE
A shared client library handling retries and timeouts.
WHAT TO ADOPT FIRST IF YOU DO WANT ONE
Encryption and metrics.
WHY
They deliver most of the value with least configuration.
WHAT TO BE CAUTIOUS OF
Traffic policies nobody can reason about.
WHAT TO ESTABLISH BEFORE ADOPTING
Who will operate it.
