Shilin He

Effective incident management in large-scale IT systems relies on troubleshooting guides (TSGs), but their manual execution is slow and error-prone. While recent advances in LLMs offer promise for automating incident management tasks, existing LLM-based solutions lack specialized support for several key challenges, including managing TSG quality issues, interpreting complex control flow, handling data-intensive queries, and exploiting execution parallelism. We first conducted an empirical study on 92 real-world TSGs, and, guided by our findings, we present , a novel end-to-end agentic framework for troubleshooting guide automation. Our approach features a three-stage workflow: the first stage provides a comprehensive guide together with a tool, TSG Mentor, to assist site reliability engineers (SREs) in improving TSG quality; the second stage performs offline preprocessing using LLMs to extract structured execution directed acyclic graphs (DAGs) from unstructured TSGs and to create dedicated Query Preparation Plugins (QPPs); and the third stage executes online using a DAG-guided scheduler-executor framework with a memory system to ensure correct workflow and support parallel execution of independent steps. Our empirical evaluation on a collection of real-world TSGs and incidents demonstrates that achieves a ∼94% success rate on GPT-4.1, outperforming baselines with less time and token consumption. Furthermore, it achieves a remarkable execution time reduction of 32.9% to 70.4% for parallelizable TSGs. Our code and sample data are publicly available at https://github.com/microsoft/StepFly.

Dynamic languages like Python are highly valued for the flexibility they offer, primarily due to their dynamic typing system. However, this flexibility often comes at a cost: type-related errors typically manifest only at runtime, which can result in elusive bugs. This scenario highlights the urgent need for effective automatic type inference strategies. Recently, there has been significant progress in the development of deep learning-based methods targeting the issues. Despite their potential, these methods face considerable limitations in the context of rare types (predominantly user-defined types in Python), which are often underrepresented in training datasets. To tackle the inherent unpredictability of dynamic characteristics in Python, we present an innovative framework named RunTyper. RunTyper uniquely integrates dynamic analysis with a sophisticated framework that combines dynamic instrumentation and static rules. This hybrid approach is utilized to validate and refine predictions made by deep neural networks. When compared with Type4Py, a deep learning-based type inference tool, the experiment results indicate that RunTyper achieves a significantly higher accuracy rate of 53.7%, whereas Type4Py attains only 41.9%. Additionally, a series of experiments demonstrate that RunTyper significantly enhances performance in user-defined types. When compared to HiTyper, which similarly seeks to refine deep type inference but is limited to static rules, RunTyper exhibits a notable improvement, increasing accuracy from 17.0% to 22.9%, an enhancement rate of 35%. Moreover, the experiments indicate that RunTyper improves the accuracy of GPT-4 in processing user-defined types by a remarkable 17%. These findings emphasize RunTyper’s proficiency in refining and enhancing the results produced by deep prediction models as well as its ability to accurately infer rare types that are traditionally difficult to predict.

Large-scale cloud systems play a pivotal role in modern IT infrastructure. However, incidents occurring within these systems can lead to service disruptions and adversely affect user experience. To swiftly resolve such incidents, on-call engineers depend on crafting domain-specific language (DSL) queries to analyze telemetry data. However, writing these queries can be challenging and time-consuming. This paper presents a thorough empirical study on the utilization of queries of KQL, a DSL employed for incident management in a large-scale cloud management system at Microsoft. The findings obtained underscore the importance and viability of KQL queries recommendation to enhance incident management. Building upon these valuable insights, we introduce Xpert, an end-to-end machine learning framework that automates KQL recommendation process. By leveraging historical incident data and large language models, Xpert generates customized KQL queries tailored to new incidents. Furthermore, Xpert incorporates a novel performance metric called Xcore, enabling a thorough evaluation of query quality from three comprehensive perspectives. We conduct extensive evaluations of Xpert, demonstrating its effectiveness in offline settings. Notably, we deploy Xpert in the real production environment of a large-scale incident management system in Microsoft, validating its efficiency in supporting incident management. To the best of our knowledge, this paper represents the first empirical study of its kind, and Xpert stands as a pioneering DSL query recommendation framework designed for incident management.

Log parsing, which extracts log templates and parameters, is a critical prerequisite step for automated log analysis techniques. Though existing log parsers have achieved promising accuracy on public log datasets, they still face many challenges when applied in the industry. Through studying the characteristics of real-world log data and analyzing the limitations of existing log parsers, we identify two problems. Firstly, it is non-trivial to scale a log parser to a vast number of logs, especially in real-world scenarios where the log data is extremely imbalanced. Secondly, existing log parsers overlook the importance of user feedback, which is imperative for parser fine-tuning under the continuous evolution of log data. To overcome the challenges, we propose SPINE, which is a highly scalable log parser with user feedback guidance. Based on our log parser equipped with initial grouping and progressive clustering, we propose a novel log data scheduling algorithm to improve the efficiency of parallelization under the large-scale imbalanced log data. Besides, we introduce user feedback to make the parser fast adapt to the evolving logs. We evaluated SPINE on 16 public log datasets. SPINE achieves more than 0.90 parsing accuracy on average with the highest parsing efficiency, which outperforms the state-of-the-art log parsers. We also evaluated SPINE in the production environment of Microsoft, in which SPINE can parse 30 million logs in less than 8 minutes under 16 executors, achieving near real-time performance. In addition, our evaluations show that SPINE can consistently achieve good accuracy under log evolution with a moderate number of user feedback.

Logging, which aims to determine the position of logging statements, the verbosity levels, and the log messages, is a crucial process for software reliability enhancement. In recent years, numerous automatic logging tools have been designed to assist developers in one of the logging tasks (e.g., providing suggestions on whether to log in try-catch blocks). These tools are useful in certain situations yet cannot provide a comprehensive logging solution in general. Moreover, although recent research has started to explore end-to-end logging, it is still largely constrained by the high cost of fine-tuning, hindering its practical usefulness in software development. To address these problems, this paper proposes UniLog, an automatic logging framework based on the in-context learning (ICL) paradigm of large language models (LLMs). Specifically, UniLog can generate an appropriate logging statement with only a prompt containing five demonstration examples without any model tuning. In addition, UniLog can further enhance its logging ability after warmup with only a few hundred random samples. We evaluated UniLog on a large dataset containing 12,012 code snippets extracted from 1,465 GitHub repositories. The results show that UniLog achieved the state-of-the-art performance in automatic logging: (1) 76.9% accuracy in selecting logging positions, (2) 72.3% accuracy in predicting verbosity levels, and (3) 27.1 BLEU-4 score in generating log messages. Meanwhile, UniLog requires less than 4% of the parameter tuning time needed by fine-tuning the same LLM.

In distributed systems and microservice applications, tracing is a crucial observability signal employed for comprehending their internal states. To mitigate the overhead associated with distributed tracing, most tracing frameworks utilize a uniform sampling strategy, which retains only a subset of traces. However, this approach is insufficient for preserving system observability. This is primarily attributed to the long-tail distribution of traces in practice, which results in the omission or rarity of minority yet critical traces after sampling. In this study, we introduce an observability-preserving trace sampling method, denoted as STEAM, which aims to retain as much information as possible in the sampled traces. We employ Graph Neural Networks (GNN) for trace representation, while incorporating domain knowledge of trace comparison through logical clauses. Subsequently, we employ a scalable approach to sample traces, emphasizing mutually dissimilar traces. STEAM has been implemented on top of OpenTelemetry, comprising approximately 1.6K lines of Golang code and 2K lines of Python code. Evaluation on four benchmark microservice applications and a production system demonstrates the superior performance of our approach compared to baseline methods. Furthermore, STEAM is capable of processing 15,000 traces in approximately 4 seconds.

Cloud systems have become increasingly popular in recent years due to their flexibility and scalability. Each time cloud computing applications and services hosted on the cloud are affected by a cloud outage, users can experience slow response times, connection issues or total service disruption, resulting in a significant negative business impact. Outages are usually comprised of several concurring events/source causes, and therefore understanding the context of outages is a very challenging yet crucial first step toward mitigating and resolving outages. In current practice, on-call engineers with in-depth domain knowledge, have to manually assess and summarize outages when they happen, which is time-consuming and labor-intensive. In this paper, we first present a large-scale empirical study investigating the way on-call engineers currently deal with cloud outages at Microsoft, and then present and empirically validate a novel approach (dubbed Oasis) to help the engineers in this task. Oasis is able to automatically assess the impact scope of outages as well as to produce human-readable summarization. Specifically, Oasis first assesses the impact scope of an outage by aggregating relevant incidents via multiple techniques. Then, it generates a human-readable summary by leveraging fine-tuned large language models like GPT-3.x. The impact assessment component of Oasis was introduced in Microsoft over three years ago, and it is now widely adopted, while the outage summarization component has been recently introduced, and in this article we present the results of an empirical evaluation we carried out on 18 real-world cloud systems as well as a human-based evaluation with outage owners. The results obtained show that Oasis can effectively and efficiently summarize outages, and lead Microsoft to deploy its first prototype which is currently under experimental adoption by some of the incident teams.