How to use the AIOps Solutions Expert Service - Support - Alibaba Cloud Documentation Center

1. Service overview

1.1. Service description

As cloud computing adoption grows, the focus has shifted from migrating to the cloud to using the cloud effectively. In complex cloud or multicloud environments, O&M engineers often face challenges such as complex technology stacks, time-consuming alert configuration, missed metric-based alerting, alert storms, and slow fault localization. These issues can lead to serious financial losses.

Artificial intelligence for IT operations (AIOps) applies AI to O&M by using machine learning to improve efficiency. AIOps provides capabilities such as time series metric anomaly diagnosis, root cause analysis, resource orchestration, and automatic fault recovery. A common AIOps workflow includes real-time anomaly detection for key performance indicators (KPIs), root cause analysis for multi-dimensional metrics to quickly identify abnormal dimensions, and application topology-based root cause localization. It also uses a Configuration Management Database (CMDB) and association analysis to build root cause context and help you fix problems quickly.

The AIOps Solutions Expert Service provides solution support and consulting for your AIOps technical needs. Built on Alibaba's extensive AI experience and Alibaba Cloud specialists' expertise, it uses algorithms to model and analyze monitoring metrics in real time. When a business anomaly occurs, the service performs real-time alert association analysis and convergence, helping you reduce the mean time to repair (MTTR) and improve business stability.

The service covers time series forecasting, risk and anomaly inspection, intelligent AI monitoring and diagnostics, and intelligent root cause recommendations. It addresses fragmented O&M systems, traditional O&M methods, low O&M efficiency, and low resource utilization by integrating full-stack monitoring from infrastructure to applications. It also provides intelligent O&M capabilities for scene orchestration to facilitate automated management. The scope of specific projects can be adjusted to meet your requirements.

2. Service scope

2.1. Consulting Edition service scope

One consulting package includes 10 consecutive business days of remote AIOps solution design services. The service includes the following:

Business architecture survey

Surveys, interviews, and other methods are used to investigate and analyze the current state of your application technology stack, resource usage, and AIOps feasibility. This helps determine the number, priority, and strategy for your business systems. Based on the assessment results, technology selection recommendations for the cloud are provided.

Intelligent fault discovery solution

Based on the survey and assessment results, an intelligent fault discovery solution is designed for you. This includes unified monitoring data ingestion from multiple accounts and AI algorithm capabilities based on application groups for real-time fault discovery. Based on the analysis, the root cause is intelligently located and a stability solution with real-time anomaly detection is provided to ensure your business stability.

Customized business risk inspection solution

Based on Alibaba Cloud's business resources and technical capabilities, a customized business risk inspection solution is designed for you. This includes a customized architecture for business risk inspection capabilities and a specific implementation plan for business risk scenarios.

This service does not include the following:

The solution design focuses on the technical component layer. It does not perform a detailed analysis of your business or design application migration plans for each system individually.
The consulting service provides technical solution design based on your current system on Alibaba Cloud. It does not include overall cloud architecture consulting. If you require cloud architecture consulting, you must purchase that service separately. This service also does not include code development or diagnosis.
The service delivery method is not restricted. Alibaba Cloud conducts detailed surveys and designs consulting solutions on-site or remotely based on project needs to produce the final results.
Alibaba Cloud is not responsible for providing technical documents other than official Alibaba Cloud documents and the solution designs for intelligent fault discovery and customized business risk inspection.
Alibaba Cloud is not responsible for any implementation or maintenance of your business system planning, architecture design, or application processes.
Alibaba Cloud is not responsible for handling issues, providing technical support, or answering questions related to non-Alibaba Cloud platforms, such as third-party software and application systems.
Specific implementation work after the solution is completed is not within the scope of the consulting service.

2.2. Basic Edition service scope

One Basic package includes 10 consecutive business days of remote AIOps solution implementation assistance. Alibaba Cloud will provide up to one on-site support visit of no more than two person-days if required by the project.
Based on the consultation results, we help you select an AIOps solution. The process includes the following:
- Assist with activating cloud resources and Alibaba Cloud accounts, and completing the intelligent fault discovery and customized business risk inspection solutions.
- Provide technical support, troubleshooting, and solutions for problems encountered with the solution itself during implementation.
- Assist with application integration to help you quickly integrate applications into the AIOps solution product.

This service does not include the following:

This service does not guarantee any deliverables. The service is limited to the service period and terminates when the service period expires.
During the service, Alibaba Cloud is not responsible for implementation tasks such as deploying your applications, modifying application or data code, or performing data migration. You are responsible for all implementation work. During the implementation process, Alibaba Cloud provides technical support, guidance, and assistance with resolving issues related to Alibaba Cloud products.
Party B is not responsible for the project schedule and is not liable for any delays caused by Party A.

2.3. Standard Edition service scope

One Standard package includes 10 consecutive business days of on-site implementation assistance.
Based on the results of the consulting service, we help you implement the AIOps solution. The work includes the following:
- Assist with activating cloud resources and Alibaba Cloud accounts, and completing your intelligent fault discovery solution and customized business risk inspection solution.
- Provide technical support, troubleshooting, and solutions for problems encountered with the solution itself during implementation.
- Assist with application integration to help you quickly integrate applications into the AIOps solution product.
- You can separately purchase an offline output platform and a visualization platform to assist with on-premises deployment.

This service does not include the following:

This service does not guarantee any deliverables. The service is limited to the service period and terminates when the service period expires.
During the service, Alibaba Cloud is not responsible for implementation tasks such as deploying your applications, modifying application or data code, or performing data migration. You are responsible for all implementation work. During the implementation process, Alibaba Cloud provides technical support, guidance, and assistance with resolving issues related to Alibaba Cloud products.
Alibaba Cloud is not responsible for the project schedule. If the project is delayed for reasons attributable to you, Alibaba Cloud is not liable.

3. Prerequisites

You must request this service at least 15 business days in advance. This allows Alibaba Cloud to assess the feasibility of your business goals and timeline and confirm acceptance of the service request.
You must promptly provide Alibaba Cloud with all necessary documents, information, data, charts, system permissions, and remote access channels required for service delivery. All information provided is subject to the confidentiality clauses of this agreement. You confirm that all information you disclose to Alibaba Cloud is true, accurate, and not misleading.
The location for delivering consulting services is not restricted. Services are delivered primarily by phone, DingTalk, and email.
You are responsible for implementation during project delivery. Alibaba Cloud provides solution design and troubleshooting support. You must perform all application deployment and testing.
Both parties will use mutually agreed-upon communication methods during the project. The project managers for both parties are responsible for transmitting all written information required for the project. Communication methods may include DingTalk, the Internet, and email.
All project deliverables are in Simplified Chinese. The working language is Chinese. All deliverables are provided in Microsoft Office formats, including PowerPoint, Word, Excel, and Visio, and are submitted as electronic copies.
Party A and Party B shall perform the project work according to the agreed-upon work plan, personnel resource plan, and the start and end dates confirmed in writing. If the launch of an iteration for Party A's related business systems is delayed, the project schedule will be extended accordingly. Party B shall not be held responsible for such a delay.
Neither party is liable for any special, incidental, or indirect damages or consequential economic losses, including lost profits or savings, under this project, even if advised of the possibility of such damages.
As the operations and maintenance (O&M) entity, you are responsible for all O&M work related to your business.

4. Division of responsibilities

4.1. Customer and Alibaba Cloud

Both parties agree on and confirm the specific business goals and scope.
The contract is signed and payment is completed.

4.2. Responsibilities of both parties in each phase

Both parties have distinct responsibilities in each phase to ensure smooth project progression.

Project phase

Customer

Party B

Project preparation phase

1. Appoint a project manager with appropriate skills and experience as the main contact for communication with Alibaba Cloud. This person will be directly responsible for planning, coordinating, supervising, and controlling the project implementation, along with escalating issues and risks. They will also have full authority to make decisions on behalf of the customer in all aspects of this project.

2. The project manager must work with Alibaba Cloud personnel to confirm all matters in the project preparation phase. For more information, see Section 3: Prerequisites.

3. Prepare the office environment and approve access permissions for personnel.

4. Manage the customer's project team, communicate with personnel involved in each phase, and secure their commitment and time for the project.

5. Manage the customer's project team personnel.

1. Appoint an experienced project manager to perform project management, introduce and manage Alibaba Cloud's project team members, and communicate with the customer's project manager.

2. Propose solutions and plans for all matters in the project preparation phase and confirm them in writing with the customer's project manager. For more information, see Section 3: Prerequisites.

Current state analysis phase

1. Organize key users to participate in survey interviews according to the project plan and interview plan.

2. Provide an overview of the existing business, such as systems, applications, data, organization chart, and personnel division, as input for Alibaba Cloud's assessment based on the survey requirements.

3. Confirm Alibaba Cloud's AIOps strategy and risk governance policy.

4. Appoint a representative to review Alibaba Cloud's deliverables and provide feedback and acceptance confirmation.

1. Provide an interview plan and assess the existing infrastructure, application architecture, and application dependencies based on the interview results.

2. Propose an AIOps policy and risk governance policy based on the assessment results and obtain agreement from the customer.

3. Ensure that the final deliverables meet the acceptance standards based on the customer's acceptance feedback.

Solution design phase

1. Collaborate to design the AIOps solution.

2. Overall solution design

3. Appoint a representative to review Alibaba Cloud's deliverables and provide feedback and acceptance confirmation.

1. Design the AIOps solution based on the service scope and the customer's specific scenarios.

2. Ensure that the final deliverables meet the acceptance standards based on the customer's acceptance feedback.

Solution implementation assistance phase

1. Provide the necessary business inputs, resources, and environments to help you validate the solution's feasibility,

1. Collaborate on specific code modifications and their implementation. 2. Designate an acceptance reviewer to review deliverables, provide feedback, and confirm acceptance.

1. Assist and guide the customer in activating or purchasing cloud resources and completing the basic architecture setup and configuration.

2. Provide implementation support, troubleshooting, and solutions for the AIOps solution.

3. Build a validation demo according to the AIOps solution.

4. Provide after-sales training services according to AIOps standards.

Note: The Consulting Edition includes the preparation, analysis, and design phases. The Basic and Standard Editions include the preparation, analysis, design, and implementation phases.

5. Service Catalog

The AIOps Solutions Expert Service includes the following services:

Phase name	Service Catalog	Consulting Service	Basic Edition Service	Standard Edition Service
Current state analysis	System survey and assessment	Support
Current state analysis	Solution communication and planning	Supported
Solution design	Intelligent fault discovery solution	Support
Solution design	Customized business risk inspection solution	Support
AIOps solution implementation support	Intelligent fault discovery solution implementation support		Support	Support
AIOps solution implementation support	Customized business risk inspection solution implementation support		Support	Support
AIOps solution on-site deployment	Intelligent fault discovery solution offline output			Supported
AIOps solution on-site deployment	Customized business risk inspection solution offline output			Supported

Note: The service content for the Basic and Standard Editions is mostly the same. The Basic Edition provides remote support, whereas the Standard Edition provides on-site support and allows for the separate purchase of an offline output solution.

5.1. Service content

AIOps Solutions Expert Service
No.	Service category	Service description	Deliverable
1	Business architecture survey	Surveys the resources, business status, and core logic of your cloud-based application systems. The survey scope includes basic resource survey, business status survey, and application system survey.	Survey Report
2	Intelligent fault discovery solution design	Establishes business group units based on business data and resource group dimensions. The metrics of these units are analyzed in real time using intelligent AI algorithms to discover faults, list suspicious root cause events, and rank them by suspicion level. Based on intelligent AI algorithms and Alibaba's AIOps experience, an automated solution is provided after a fault occurs. The solution design covers three major algorithm scenarios: time series forecasting, root cause analysis, and historical data prediction. For more information about specific algorithms, see Section 10.1: List of intelligent algorithms.	Intelligent Fault Discovery Design Solution
3	Customized business risk inspection solution design	Based on Alibaba Cloud's business resources and technical capabilities, a customized business risk inspection capability architecture is designed for your business, along with a specific implementation plan for business risk scenarios. Two E-commerce application risk inspection scenarios are also provided to help you implement custom risk scenarios. For more information, see Section 10.2: List of business risk scenarios.	Customized Business Risk Inspection Design Solution

AIOps Solutions Basic Edition Service
No.	Service category	Service description	Deliverable
4	Intelligent fault discovery solution implementation	Provides an implementation plan based on the consulting service design solution to help you quickly integrate applications into the AIOps solution product.	Intelligent AI Detection Implementation Plan
5	Customized business risk inspection solution implementation	Provides an implementation plan based on the consulting service design solution to help you quickly integrate applications into the AIOps solution product.	Customized Business Risk Inspection Implementation Plan

AIOps Solutions Standard Edition Service
No.	Service category	Service description	Deliverable
6	Intelligent fault discovery on-site implementation	Provides an implementation plan based on the consulting service to help you quickly integrate applications into the AIOps solution product.	Intelligent AI Detection Implementation Plan
7	Customized business risk inspection on-site implementation	Provides an implementation plan based on the consulting service to help you quickly integrate applications into the AIOps solution product.	Customized Business Risk Inspection Implementation Plan

6. Service SLA

The SLA for the AIOps Solutions Consulting Service includes the following:

Providing AIOps solutions consulting services.
Providing a support group and on-site support as needed during the service period.
Providing solutions such as the Customized Business Risk Inspection Solution and the Intelligent Fault Discovery Solution according to the corresponding service specifications. Specific deliverables vary depending on the service content.

7. Service flow

Application time limit: You must submit a service purchase request at least 15 days before the service starts.

AIOps Solutions Consulting Service flow:

AIOps Solutions Implementation Service flow:

8. Acceptance criteria

8.1. Acceptance checklist

No.	Delivery phase	Deliverable	Deliverable type
1	Project survey phase	Project Survey Report	Document
2	Solution design phase	Intelligent Fault Discovery Solution	Support
3	Solution design phase	Customized Business Risk Inspection Solution	Document
4	Solution implementation phase	Intelligent Fault Discovery Solution Implementation Plan	Document
5	Solution implementation phase	Customized Business Risk Inspection Implementation Plan	Document

8.2. Acceptance criteria

Acceptance criteria
- The solution design meets your requirements and is confirmed online. For more information about the solution, see Section 8.1: Acceptance checklist.
- Alibaba Cloud provides the Intelligent Fault Discovery Solution and the Customized Business Risk Inspection Solution during the project. After you accept the design solution, the first phase of work is complete. The second phase, implementation, begins based on this approved solution. Alibaba Cloud provides an implementation assistance plan and technical support to help you implement the solution. After implementation assistance is complete, you must confirm the solution within five business days and complete the acceptance by signing the Service Acceptance Report online.

8.3. Acceptance plan

Party A agrees to accept the deliverables from Party B according to the following acceptance plan, which is based on the content and deliverables for each phase listed in the '8.1 Itemized Acceptance Checklist'.

No.	Acceptance milestone	Acceptance content	Acceptance completion flag
1	AIOps solution design and validation complete	All deliverables from the project preparation, survey and assessment, and design phases	Customer confirms the Acceptance Report online
2	AIOps solution implementation complete	All deliverables from the implementation phase	Customer confirms the Acceptance Report online

9. Completion flag

The customer acceptance process is complete.

10. Appendix

10.1. List of intelligent algorithms

Type	Algorithm name	Algorithm logic
Anomaly diagnosis algorithm	One-Class SVM	Performs algorithm learning and anomaly diagnosis based on historical batch data.
Anomaly diagnosis algorithm	Isolation Forest	Performs algorithm learning and anomaly diagnosis based on historical batch data.
Anomaly diagnosis algorithm	Robust Covariance	Performs algorithm learning and anomaly diagnosis based on historical batch data.
Anomaly diagnosis algorithm	LocalOutlierFactor	Performs algorithm learning and anomaly diagnosis based on historical batch data.
Anomaly diagnosis algorithm	AutoEncoder	Performs algorithm learning and anomaly diagnosis based on historical batch data.
Root cause analysis algorithm	Random Forest and PCA	Root cause analysis
Time series anomaly diagnosis algorithm	K-Sigma	Performs anomaly diagnosis based on real-time series data.
Time series anomaly diagnosis algorithm	ARIMA	Performs anomaly diagnosis based on real-time series data.

10.2. List of business risk scenarios

Scenario category	Scenario name	Description
E-commerce	Create fulfillment order	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Product page	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Add to shopping cart	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Shopping cart rendering	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Order page rendering	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Submit for payment	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Payment result	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Checkout page rendering	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.
E-commerce	Create fulfillment order	Calculate scenario metrics from scenario logs to quickly implement customized business scenario inspections.

10.3. Expected solution outcomes

Intelligent fault discovery solution

The solution includes link-based root cause analysis for multiple metrics and products. It provides eight real-time anomaly detection and root cause analysis algorithm scenarios while maintaining algorithm versatility.

Customized business risk solution

The solution includes customized inspections for various business scenarios. It provides a scenario architecture plan and two E-commerce application risk inspection scenarios to help you implement custom risk scenarios.