distributed-audit

A distributed systems correctness audit identifies concurrency bugs, replication correctness, and failover safety issues by analyzing risk domains like shard-local concurrency and cluster topology. It evaluates fault scenarios including timing faults, crash faults, and network faults.

To test failover safety and cross-shard atomicity, run a correctness audit that evaluates these specific risk domains against potential fault scenarios. The audit then recommends existing test suites and defines new tests to validate cluster topology resilience.

The best way to check for network and protocol faults is to perform a structured correctness audit using the Full Report tier. This evaluates fault scenarios like resource exhaustion and protocol failures, outputting detailed audit templates for your system.

Yes, you can audit design documentation instead of source code using the Design Review Mode. This mode provides guidance for evaluating design docs against invariants, fault taxonomy, and testing patterns without requiring code analysis.

Yes, distributed systems correctness audits support quick assessments of small changes through the Quick Assessment tier. This provides a lightweight output template to rapidly evaluate specific risk domains without generating a full report.

Before running a distributed systems correctness audit, you need in-depth knowledge of distributed system concepts and practices. Understanding risk domains like client-observable consistency, pub/sub routing, and persistence is required to interpret audit results.