Skip to main content

Setting Up WME AI Observability

WME AI Observability is a monitoring and analytics service that provides visibility into your AI workloads running on the WaveMaker Enterprise platform. It is built on Langfuse for LLM tracing and a custom Analytics service. Install and run this service on the WME AI Observability instance.

Step 1 — Download and Extract Package

Download the Package

Download the package using the link shared by the WaveMaker team:

wget https://s3.us-east-1.amazonaws.com/downloads.wavemaker.com/wme/<WME-AI-Version>/rhel/ai-observability-<WME-AI-Version>.tar.gz

Create Directory

mkdir ai-observability

Extract the Package

tar -xzf ai-observability-<WME-AI-Version>.tar.gz -C ai-observability

Step 2 — Package Structure

The extracted package contains installation scripts, Docker images, and configuration files.

Directory Structure

ai-observability-<WME-AI-Version>/
├── INSTALL.md
├── certs/
├── docker-compose.yml
├── env.example
├── images.tar.gz
├── init-db.sql
├── install-ai-observability.sh
└── nginx.conf.template

Step 3 — Installation Scenarios

Case 1 — Installation Without Public Domain

Use this method when accessing the application using:

http://<Public-IP>:5050
caution

This method uses unencrypted HTTP access. It is recommended only for internal or non-production environments. Use Case 2 for secure HTTPS access.

Run the Installation Script

Navigate into the extracted directory and run the installation script:

cd ai-observability/ai-observability-<WME-AI-Version>
bash install-ai-observability.sh

During execution, provide the following inputs:

PromptValue
Public domain for Analytics UI [analytics.example.com]:Press Enter
Run nginx reverse proxy (HTTPS on 443) for analytics.example.com? [Y/n]n
WaveMaker platform instance IP (leave blank to skip Consul registration) [127.0.0.1]Platform Private IP
Private IP for Langfuse (registered in Consul; usually this host)Private IP of WME AI Observability
Langfuse project PUBLIC key (blank = auto-generate)Press Enter
Langfuse project SECRET key (blank = auto-generate)Press Enter
Analytics API key (blank = auto-generate)Press Enter

Accessing the Application

After successful installation, access the application at:

http://<Public-IP-of-WME-AI-Observability>:5050

Login Password

The login password is the generated ANALYTICS_API_KEY. You can find this value in the terminal output after execution of install-ai-observability.sh.

Case 2 — Installation With Public Domain and SSL

Use this method when accessing the application through a domain with HTTPS enabled.

SSL Certificate Preparation

Place the SSL certificate files inside:

ai-observability/ai-observability-<WME-AI-Version>/certs/

Required file names:

FileRequired Name
Certificate Filebundle.crt
Private Key Fileprivate.key

Directory example:

certs/
├── bundle.crt
└── private.key

Run the Installation Script

Navigate into the extracted directory and run the installation script:

cd ai-observability/ai-observability-<WME-AI-Version>
bash install-ai-observability.sh

During execution, provide the following inputs:

PromptValue
Public domain for Analytics UI [analytics.example.com]:<domain-name>
Run nginx reverse proxy (HTTPS on 443) for <domain-name>? [Y/n]Y
Also expose Langfuse UI via nginx? (otherwise internal only) [y/N]y
Domain for Langfuse [<domain-name>]Press Enter
Also expose MinIO API via nginx? (otherwise internal only) [y/N]y
Domain for MinIO [minio.<domain-name>]Press Enter
WaveMaker platform instance IP (leave blank to skip Consul registration) [127.0.0.1]Platform Private IP
Langfuse project PUBLIC key (blank = auto-generate)Press Enter
Langfuse project SECRET key (blank = auto-generate)Press Enter
Analytics API key (blank = auto-generate)Press Enter
TLS certs directory [./certs]Press Enter
note

By pressing Enter for TLS certs directory [./certs], the installer automatically uses:

  • ai-observability/ai-observability-<WME-AI-Version>/certs/bundle.crt
  • ai-observability/ai-observability-<WME-AI-Version>/certs/private.key

Accessing the Application

After successful installation, access the application at:

https://wmai-analytics.[mycompany].com

The Analytics URL will also be displayed in the terminal output after execution of install-ai-observability.sh.

Login Password

The login password is the generated ANALYTICS_API_KEY. You can find this value in the terminal output after installation completes.