A pizza ordering metaphor illustrates the difference between synchronous blocking communication like a phone call and asynchronous non-blocking communication like an email.
Long-running processes are defined by periods of waiting, driven by business needs like human tasks, response delays, or mitigating buyer's remorse.
Implementing stateful, long-running logic with simple database columns leads to accidental complexity around visibility, escalations, versioning, and reliability.
A practical demonstration shows how to use a workflow engine like Camunda with Spring Boot to model, execute, and monitor a customer onboarding process.
Failures in distributed systems, such as an unavailable service during an airline check-in, necessitate long-running retry patterns to create a resilient architecture.
Safely retrying operations in a distributed system requires idempotent APIs to prevent duplicate actions and compensation logic to clean up after ultimate failures.
Empowering individual services to handle their own long-running logic, like payment retries, leads to cleaner service boundaries and avoids creating monolithic god services.
Graphical models like BPMN serve as living documentation, facilitate communication between business and technical teams, and provide crucial operational visibility.
The speaker answers audience questions about Camunda's open-source model, the trade-offs of adopting a workflow engine, and how the technology scales.
50:06Jens Happe
29:32Maxim Fateev
55:30Christian Woerz
58:05Michael Eder & Philipp Frauenthaler
58:23Luka Samaržija & Ivan Sokol
58:52Hans Hosea Schäfer
46:24David Leitner
1:01:30Henning Schwentner
.gif?w=240&auto=compress,format)
Jobs that call for the skills explored in this talk.
Saby New Compy
Bielefeld, Germany
ATR Software GmbH
Neu-Ulm, Germany
wunschlösung GmbH
Jena, Germany
wunschlösung GmbH
Jena, Germany
ATR Software GmbH
Neu-Ulm, Germany
IT-Objects GmbH
Essen, Germany