Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Model Discovery and Graph Simulation: A Lightweight Gateway to Chaos Engineering

Anatoly A. Krasnovsky

TL;DR

The paper advocates a lightweight, automated discovery of a connectivity-only service-dependency topology with replica counts to estimate endpoint availability under fail-stop faults. It demonstrates that such a graph-based approach, derived from passive artifacts like traces, can closely align with live chaos-engineering outcomes, especially when replication is used. This enables efficient resilience triage and real-time posture signaling within CI/CD and observability platforms. The work frames a practical gateway to resilience testing that reduces reliance on broad, risky chaos experiments while preserving validation for critical scenarios.

Abstract

Chaos engineering reveals resilience risks but is expensive and operationally risky to run broadly and often. Model-based analyses can estimate dependability, yet in practice they are tricky to build and keep current because models are typically handcrafted. We claim that a simple connectivity-only topological model - just the service-dependency graph plus replica counts - can provide fast, low-risk availability estimates under fail-stop faults. To make this claim practical without hand-built models, we introduce model discovery: an automated step that can run in CI/CD or as an observability-platform capability, synthesizing an explicit, analyzable model from artifacts teams already have (e.g., distributed traces, service-mesh telemetry, configs/manifests) - providing an accessible gateway for teams to begin resilience testing. As a proof by instance on the DeathStarBench Social Network, we extract the dependency graph from Jaeger and estimate availability across two deployment modes and five failure rates. The discovered model closely tracks live fault-injection results; with replication, median error at mid-range failure rates is near zero, while no-replication shows signed biases consistent with excluded mechanisms. These results create two opportunities: first, to triage and reduce the scope of expensive chaos experiments in advance, and second, to generate real-time signals on the system's resilience posture as its topology evolves, preserving live validation for the most critical or ambiguous scenarios.

Model Discovery and Graph Simulation: A Lightweight Gateway to Chaos Engineering

TL;DR

The paper advocates a lightweight, automated discovery of a connectivity-only service-dependency topology with replica counts to estimate endpoint availability under fail-stop faults. It demonstrates that such a graph-based approach, derived from passive artifacts like traces, can closely align with live chaos-engineering outcomes, especially when replication is used. This enables efficient resilience triage and real-time posture signaling within CI/CD and observability platforms. The work frames a practical gateway to resilience testing that reduces reliance on broad, risky chaos experiments while preserving validation for critical scenarios.

Abstract

Chaos engineering reveals resilience risks but is expensive and operationally risky to run broadly and often. Model-based analyses can estimate dependability, yet in practice they are tricky to build and keep current because models are typically handcrafted. We claim that a simple connectivity-only topological model - just the service-dependency graph plus replica counts - can provide fast, low-risk availability estimates under fail-stop faults. To make this claim practical without hand-built models, we introduce model discovery: an automated step that can run in CI/CD or as an observability-platform capability, synthesizing an explicit, analyzable model from artifacts teams already have (e.g., distributed traces, service-mesh telemetry, configs/manifests) - providing an accessible gateway for teams to begin resilience testing. As a proof by instance on the DeathStarBench Social Network, we extract the dependency graph from Jaeger and estimate availability across two deployment modes and five failure rates. The discovered model closely tracks live fault-injection results; with replication, median error at mid-range failure rates is near zero, while no-replication shows signed biases consistent with excluded mechanisms. These results create two opportunities: first, to triage and reduce the scope of expensive chaos experiments in advance, and second, to generate real-time signals on the system's resilience posture as its topology evolves, preserving live validation for the most critical or ambiguous scenarios.

Paper Structure

This paper contains 9 sections, 1 equation, 2 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Approach
  3. Model
  4. Discovery pipeline
  5. Availability estimation
  6. System and Experimental Setup
  7. Results
  8. Discussion and Related Work
  9. Future Plans

Figures (2)

  • Figure 1: $R_{\text{model}}$ vs. $R_{\text{live}}$ across all conditions; dashed $y{=}x$ shows ideal agreement. Replication effect captured: replicated points lie higher and closer to the diagonal than no‑replication. Points are CI‑job aggregates (4.5M simulations $+$ 450 live windows per job).
  • Figure 2: Error percentage ($\Delta\%$) by failure rate and deployment mode. Summarizes the relative model--live discrepancy per condition using the same CI-job aggregates as Fig. 1 and Table 1.