Failure Propagation and Retry Dependency Reasoning Dataset

#Code Reasoning #Dependency Reasoning #Failure Propagation Analysis #Execution Outcome Prediction #Distributed Job Scheduling #Workflow Orchestration #Fault Diagnosis #Task Recovery

3540
0.2GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-05

AI Analysis & Value Prop

This dataset focuses on distributed task scheduling and workflow orchestration, with scheduler source code that captures task dependencies and failure-handling logic, including failures, retries, skips, timeouts, cancellations, and partial success. Structured annotations cover task states, dependency graphs, failure propagation, retry decisions, execution feasibility, outcome predictions, and reasoning traces, helping models analyze how faults affect downstream tasks. It is suited to code SFT, scheduler logic analysis, failure recovery decisions, and execution outcome prediction.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
tasks	array	Tasks extracted from the source code, including their current states, dependencies, and retry configuration.
reasoning	array	Step-by-step explanation of state propagation and execution outcome judgments based on source-code control flow.
source_code	string	Original scheduler source code containing task dependencies and logic for failure handling, retries, skips, timeouts, cancellations, and partial success.
retry_decisions	array	Reasoning results for whether failed or timed-out tasks can be retried, how many retries are allowed, and what actions follow.
source_language	string	Programming language identified from source syntax and file characteristics; marked as undetermined when it cannot be identified.
analysis_summary	string	A concise summary of how failure states propagate through the dependency graph and which tasks can execute, must wait, or must terminate.
dependency_graph	object	A scheduling dependency graph composed of task nodes and directed dependency edges.
failure_propagation	array	Describes how failures, timeouts, cancellations, or partial-success states propagate through the dependency graph and affect downstream tasks.
execution_prediction	object	Prediction of the overall scheduling outcome based on the dependency graph, failure propagation, and retry policy.
execution_classification	object	Classifies downstream tasks as executable, waiting, or terminated based on failure propagation and dependency satisfaction.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: The dataset is organized around scheduler source code and task dependency logic. It covers task states, dependency graphs, failure propagation, retry decisions, execution feasibility classifications, outcome predictions, and reasoning traces. Relevant states include failure, retry, skip, timeout, cancellation, and partial success.

Does the dataset specify its size or a fixed programming language?: The available information does not specify a sample count or a fixed programming language. Each record includes a source-language field that identifies the language from syntax and file characteristics, or marks it as undetermined when it cannot be identified. Refer to the product listing for the specifications of the available release.

Which training or evaluation capabilities does it support?: It supports code SFT and code reasoning, with a focus on understanding task dependencies, tracing failure propagation, classifying downstream tasks as executable, waiting, or terminated, and predicting overall scheduling outcomes.

How can I obtain the dataset, and what licensing applies?: Availability, delivery format, and licensing are subject to the details on the product page or in the sales agreement. Before use, confirm that the license covers your intended activities, such as training, evaluation, commercial deployment, or redistribution.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Failure Propagation and Retry Dependency Reasoning Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/fc3ae66a1a2b908e475905a9a08f2205},
  urldate={2026-10-27},
  keywords={failure propagation dataset,task dependency reasoning,scheduler source code,distributed job scheduling,retry decisions,fault recovery analysis,code reasoning training,execution outcome prediction},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task