Tag
#fault-tolerance
4 articles
Supervision Trees for OSINT Pipelines: Let It Crash, Keep the Evidence
How Prismatic structures supervision trees so that a single flaky OSINT adapter never brings down a case. DynamicSupervisor, Task.Supervisor, and the one-for-one rule that saved production.
Circuit Breakers for Flaky OSINT Sources: Fail Fast, Recover Quietly
Some OSINT sources are up 99.9% of the time. Some are up 73%. Retrying a down source synchronously is how pipelines die. Circuit breakers turn a 30-second timeout into a 1-millisecond :circuit_open.
Self-Healing Systems: OTP Supervision in Practice
OTP supervision trees are the foundation of self-healing software. This article covers the SupervisionIntrospector, RemediationRegistry, restart strategies, cascade failure prevention, and the lessons learned system that makes the platform smarter after each fault.
Building Self-Healing OTP Supervision Trees
How Prismatic uses OTP supervision strategies, circuit breakers, and remediation registries to build systems that recover from failures automatically without human intervention.