AIOps: Real-World Challenges and Research Innovations
Abstract
AIOps is about empowering software and service engineers (e.g., developers, program managers, support engineers, site reliability engineers) to efficiently and effectively build and operate online services and Apps at scale with artificial intelligence (AI) and machine learning (ML) techniques. AIOps can help achieve higher service quality and customer satisfaction, engineering productivity boost, and cost reduction. In this technical briefing, we summarize the real-world challenges on building AIOps solutions based on our practice and experience in Microsoft, propose a roadmap of AIOps related research directions, and share a few successful AIOps solutions we have built for Microsoft service products.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
AIOps aims to move from monitoring towards observability and address this issue through Machine Learning.
“Artificial Intelligence for IT Operations (AIOps) [9] aims to move from monitoring towards observability and address this issue through Machine Learning (ML).”
From Toward Context-Aware Anomaly Detection for AIOps in Microservices Using Dynamic Knowledge Graphs (2026) · cites this paper as [9]
AIOps can autonomously learn normal baselines from massive historical data using machine learning and NLP algorithms.
“Leveraging machine learning and Natural Language Processing (NLP) algorithms, AIOps can autonomously learn normal baselines from massive historical data.”
From A Survey of Anomaly Detection and Root Cause Analysis Techniques for Artificial Intelligence for IT Operations (AIOps) in Large-Scale Software Systems (2026) · cites this paper as [1]
AIOps has emerged as a promising solution to address the management challenges of complex distributed systems.
“Artificial Intelligence for IT Operations (AIOps) has thus emerged as a promising solution [1] .”
From A Survey of Anomaly Detection and Root Cause Analysis Techniques for Artificial Intelligence for IT Operations (AIOps) in Large-Scale Software Systems (2026) · cites this paper as [1]
Sourcing high-quality labeled data is difficult, and organizational inertia may hinder the transition to AI-based operations.
“Moreover, sourcing high-quality labeled data is difficult, and organizational inertia may hinder the transition to AI-based operations [7,8].”
From Log anomaly detection in AIOps: A real-world implementation using Large Language Models (2026) · cites this paper as [7]
The citation lists haven’t been fetched yet
References 7 · Cited by 238
- Authors
- Dang, Y., Lin, Q., & Huang, P.
- Affiliations
- Microsoft (United States) (US)
- Microsoft Research Asia (China) (CN)
- Johns Hopkins University (US)
- Venue
- 2019 IEEE/ACM 41st International Conference on Software Engineering: Companion Proceedings (ICSE-Companion)
- Publisher
- IEEE
- Published
- 2019-05
- Pages
- 4-5
- ISSN
- 0270-5257 (ISSN-L)
- OpenAlex
- W2944703101
- Cited by
- 238
- References
- 7
- Citations by year
- 2026: 41 · 2025: 31 · 2024: 31 · 2023: 43 · 2022: 34 · 2021: 38 · 2020: 20
- FWCI
- 19.88
- Type
- conference paper
- Language
- en
- Open access
- closed
- Licence
- Locations
- ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Software System Performance and Reliability
- Physical Sciences › Computer Science › Information Systems › Cloud Computing and Resource Management
- Physical Sciences › Computer Science › Information Systems › Software Engineering Research
- Keywords
- AIOps, service quality, engineering productivity, cost reduction, research roadmap
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
