If you are in for a quick fix, you can also stub network calls to 3rd party services, such as an S3 upload and assume for example that a profile picture of the user got uploaded.
Luckily I work with monoliths all the time, so my docker compose in CI got a postgres connection, nothing else 😀
Strong take on dependency management. The Docker Compose pain is real but most teams treat it as inevitable rather than a design signal. That shift from synchronous REST to async Kafka communication is underrated becuase it flips the entire test harness from needing all services running to just needing one message broker. The autonomy loss point in #6 hits hard tho, coordinating releases across team boundaries is where velocity actually dies.
If you are in for a quick fix, you can also stub network calls to 3rd party services, such as an S3 upload and assume for example that a profile picture of the user got uploaded.
Luckily I work with monoliths all the time, so my docker compose in CI got a postgres connection, nothing else 😀
Strong take on dependency management. The Docker Compose pain is real but most teams treat it as inevitable rather than a design signal. That shift from synchronous REST to async Kafka communication is underrated becuase it flips the entire test harness from needing all services running to just needing one message broker. The autonomy loss point in #6 hits hard tho, coordinating releases across team boundaries is where velocity actually dies.
Thanks for sharing!
Agreed, coordinating different teams for a release is a real pain.