Log-Intelligence-AI-Deploy

AI Log Intelligence CLI tool for L3 support log triage: parse logs, mask sensitive values, summarize recurring errors and request AI-assisted root-cause hints.

Build

Windows

.\mvnw.cmd clean package

Unix / Linux / macOS

./mvnw clean package

Run

Windows

java -jar log-intelligence-ai-deploy.jar --inputLog=sample.out --maskConfig=mask.properties --aiPrompt=prompt.txt

Unix / Linux / macOS

java -jar log-intelligence-ai-deploy.jar --inputLog=sample.out --maskConfig=mask.properties --aiPrompt=prompt.txt

Required option:

• --inputLog=<path>

Optional options:

• --maskConfig=<path to .properties>
• --aiPrompt=<path to .txt>
• --aiMaxErrors=<number> (> 0: top N errors, <= 0: analyze all errors, default: 5)
• --skipAiModelCheck=true (skip Ollama model validation via /api/tags at startup)

Masking behavior

• The generated log summary and report keep the raw log values.
• Masking is applied only to the prompt sent to the AI model.
• AI prompt masking is controlled by src/main/resources/application.yml defaults and optional --maskConfig overrides.

Supported mask keys in mask.properties:

• mask.ip_port
• mask.email
• mask.uuid
• mask.mac
• mask.jdbc
• mask.host
• mask.cn
• mask.id
• mask.fingerprint

Example mask.properties:

mask.ip_port=true
mask.email=true
mask.uuid=true
mask.mac=true
mask.jdbc=true
mask.host=true
mask.cn=true
mask.id=true
mask.fingerprint=true

AI settings in src/main/resources/application.yml:

• spring.ai.ollama.base-url
• spring.ai.ollama.chat.options.model
• Optional AI masking defaults: logai.mask.*
• Optional extra system guidance appended to baseline guardrails: logai.ai.system-prompt

The tool writes a sidecar report file at <inputLog>.analysis.txt.