ChatOps4Msa
1.0.0
ChatOps for Microservices Architecture - ChatOps4Msa
ChatOps4Msa is a Conversational Assistant Tool designed for Microservices Development and Operations, utilizing Large Language Models (LLMs). Users can monitor real-time service status, perform automated actions like tests and code scanning, and receive alert notifications using a designed query language called CQL (ChatOps Query Language) or natural language.
Developers can customize desired functionalities using Low-code configuration files, which distinguishes it from traditional chatbot development that often requires extensive training data. Leveraging state-of-the-art Large Language Models, ChatOps4Msa analyzes user queries to determine intent and entities, enabling corresponding actions.
Here is the scenario of ChatOps4Msa:
Here is the operation concept of ChatOps4Msa:
For more details, see feature.md
In the chatops channel, use @ChatOps4Msa-Bot to mention ChatOps4Msa, then input your request in natural language.
In the chatops channel, directly execute commands using slash commands that start with /.
And here is the syntax of CQL:
/[ACTION] [TOOL] [COMMAND] [SERVICE] [FLAGS]
| PARAMETER | DESCRIPTION | REQUIRED? |
|---|---|---|
| [ACTION] | The actions to be performed, such as retrieving information (get), configuring settings (set), monitoring services (monitor), testing services (test), and so on. | required |
| [TOOL] | The tools to be used, such as GitHub, Prometheus, k6, custom modules, and so on. | required |
| [COMMAND] | The operations to be executed, such as service_recent_activity. |
required |
| [SERVICE] | The target service for execution, which can be selected as all_service to include all services. |
optional |
| [FLAGS] | Additional parameter options, which can be multiple; all CQL commands include the subscribe parameter for scheduling periodic executions of the function, with the parameter value being a Cron expression. |
optional |
Here are several demonstration capabilities designed to showcase the features of ChatOps4Msa, you can perform these capabilities using natural language or CQL.
| INTENT | ENTITY | DESCRIPTION |
|---|---|---|
| get-example-echo_message | message | Just echo a message to Discord. |
| get-example-math_calculate | expression | Perform calculations using the given mathematical expression. |
| get-github-service_recent_activity | service_name, number_of_activity | Retrieve recent activities of a service's repo. |
| get-github-service_past_week_team_activity | service_name | Get the team's activity for the specified service within the past week. |
| get-github-service_code_scanning | service_name | Get the check report of the newest commit in the service's repo. |
| set-github-issue | service_name, issue_title, issue_message | Create issues in the repo of services with dependency risks. |
| get-github-issue_progress_timeline | service_name, number_of_issue | Describe the progress of the issue with a timeline. |
| get-github-contribution_of_team_member | service_name | Get the contribution analysis of team members. |
| monitor-prometheus-promql_overview | (null) |
Monitor all service metrics using PromQL. |
| monitor-prometheus-grafana_overview | (null) |
Monitor the overall status of all services using a graphical user interface (GUI). |
| get-prometheus-alert_rule | (null) |
View the alert rules through the Prometheus UI. |
| monitor-prometheus-service_health | (null) |
Monitor the health status of all the services. |
| monitor-prometheus-all_service_yesterday_status | (null) |
Retrieve yesterday's microservice operational status using PromQL. |
| monitor-prometheus-service_dependency | (null) |
Monitor the dependency relationships between microservices. |
| test-k6-stress_testing | service_name | Perform stress testing on a specific service. |
| test-k6-smoke_testing_all_service | number_of_virtual_user | Perform smoke testing on all service. |
| test-custom_test-smoke_and_repo_all_service | number_of_virtual_user | Perform smoke tests on all services, monitor status and get recent activity for failing microservices. |
For security reasons, the following capabilities can only be executed via slash commands and not through natural language.
Fill in the Cron expression into the subscribe parameter of CQL, for example:
You can use the slash command /check_all_subscription to view all subscriptions, for example:
And you can use the slash command /unsubscribe_all_capability to unsubscribe from all capabilities, for example:
YouTube Link
We do check the health of the Discord Bot periodically. Nevertheless, we do not have high availability guarantee on the service. Feel free to open an issue if it becomes unavailable.
Discord Link
The default account for DevOps tools is
admin, and the default password issoselab.
Create a Discord Application (Bot):
Supervisor, and Team_Member.Register Slash Commands (CQL):
git clone https://github.com/sheng-kai-wang/ChatOps4Msa.gitapplication.properties and secret.yml files, you can refer to the provided templates: application-template.properties and secret_template.yml.Deploy to your server using the provided .sh script files:
sh run_all.sh to start the entire system."Ctrl" + "b", then "d". You can re-enter the session with tmux attach.Fork the following 4 microservice repositories to utilize GitHub-related capabilities:
<<Reverse proxy domain>>adminsoselabscan_run_id from the URL. Refer to the step_for_GitHub_scan_run_id.Log in to Grafana and manually configure the data source (Prometheus) to enable Grafana-related capabilities:
sh run_chatops4msa.sh to do so.Only need to write configuration files in low-code format.
See custom_capability.md
