Logara-AI

`POST /ingest` examples

Raw log input

curl -X POST http://localhost:8000/ingest \
  -H "Content-Type: application/json" \
  -d '{"log_data": "ERROR: Database timeout for user 123"}'

Example response:

{
  "status": "success_queued",
  "parsed": {
    "timestamp": null,
    "level": "ERROR",
    "service": null,
    "host": null,
    "message": "ERROR: Database timeout for user 123",
    "source": null,
    "metadata": {},
    "parser_type": "structured_input",
    "raw": "ERROR: Database timeout for user 123"
  },
  "structured_output": {
    "Problem Summary": "Detected issue in logs: ERROR: Database timeout for user 123",
    "Possible Cause": "System error, timeout, invalid request, or service failure.",
    "Affected Component": "Backend ingestion pipeline / API / service layer",
    "Suggested Fix": "Check logs, validate input, monitor service health, and retry request."
  },
  "metadata": {},
  "redaction_summary": {}
}

Structured log input

curl -X POST http://localhost:8000/ingest \
  -H "Content-Type: application/json" \
  -d '{
    "timestamp": "2026-05-26T10:00:00Z",
    "level": "ERROR",
    "service": "payments-service",
    "message": "Database timeout for user 123",
    "source": "api",
    "metadata": {
      "request_id": "req-123"
    }
  }'

Example response:

{
  "status": "success_queued",
  "parsed": {
    "timestamp": "2026-05-26T10:00:00Z",
    "level": "ERROR",
    "service": "payments-service",
    "host": null,
    "message": "Database timeout for user 123",
    "source": "api",
    "metadata": {
      "request_id": "req-123"
    },
    "parser_type": "structured_input",
    "raw": "{\"timestamp\":\"2026-05-26T10:00:00Z\",\"level\":\"ERROR\",\"service\":\"payments-service\",\"message\":\"Database timeout for user 123\",\"source\":\"api\",\"metadata\":{\"request_id\":\"req-123\"}}"
  },
  "structured_output": {
    "Problem Summary": "Detected issue in logs: Database timeout for user 123",
    "Possible Cause": "System error, timeout, invalid request, or service failure.",
    "Affected Component": "Backend ingestion pipeline / API / service layer",
    "Suggested Fix": "Check logs, validate input, monitor service health, and retry request."
  },
  "metadata": {
    "request_id": "req-123"
  },
  "redaction_summary": {}
}

`POST /v1/logs` examples

OTel log batch

curl -X POST http://localhost:8000/v1/logs \
  -H "Content-Type: application/json" \
  -d '{
    "resourceLogs": [
      {
        "scopeLogs": [
          {
            "logRecords": [
              {
                "timeUnixNano": 1762350000000000000,
                "severityNumber": 17,
                "severityText": "ERROR",
                "body": {"stringValue": "Cache miss spike detected"},
                "attributes": {
                  "service.name": "cache-service"
                }
              }
            ]
          }
        ]
      }
    ]
  }'

Example response:

{
  "status": "success",
  "message": "OTel logs processed successfully",
  "processed_records": 1,
  "redaction_summary": {},
  "fallback_used": false
}

Notes

• POST /ingest accepts either a raw log_data string or a structured JSON object.
• POST /v1/logs accepts OpenTelemetry HTTP export payloads and normalizes them before queueing.
• The worker processes queued payloads asynchronously and applies semantic duplicate clustering when enabled.

Quick Start (Local Dev)

1. Clone & Setup:

   git clone https://github.com/Dharanish-AM/Logara-AI.git
   cd Logara-AI
   

Before running, set your Redis password in .env:

```bash cp .env.example .env

Edit .env and set REDIS_PASSWORD

2. **Start Infrastructure**:

   

3. **Backend**:

   

# In terminal 1: Start the ingestor API fastapi dev main.py

# In terminal 2: Start the background log processor python worker.py

4. **AI Engine Service**:

   

# Start the AI Engine on port 8001 uvicorn main:app --port 8001

5. **Frontend**:

   

Ingestion API Endpoints

Logara AI provides two main ingestion endpoints:

1. Standard Ingest (POST /ingest): - For single, raw log strings. - Body format: {"log_data": "[2026-05-16 10:30:00] INFO: service started"}

2. OpenTelemetry Log Ingest (POST /v1/logs): - For standard OpenTelemetry (OTLP) log collector HTTP exports. - Accepts standard JSON batches of resource logs, scope logs, and log records. - Automatically merges resource attributes, extracts timestamps/severity levels, and preserves metadata.

Security & Redaction Pipeline

Logara AI includes a configurable backend redaction pipeline designed to sanitize sensitive information before logs enter queue processing, vectorization, or downstream AI workflows.