Observability

Lexia's Observability layer provides complete visibility into system health, performance, and reliability — enabling proactive monitoring, alerting, and long-term stability of your agents.

Image: Diagram showing Observability at the base of Lexia's Safety stack connected to Logs, Alerts, and Metrics

1. Overview

The Observability Layer extends Tracing and Insights with infrastructure-level monitoring.
It integrates runtime logs, metrics, and alerts to give teams real-time visibility over every component — agents, databases, and microservices.

You can connect Lexia with tools like:

Better Stack — logging and uptime monitoring
Wazuh — security and anomaly detection
Grafana / Prometheus — advanced visualization

2. Logs & Metrics

Lexia captures structured logs from every agent and service:

Request latency
Function call duration
Memory and CPU usage
Guardrail and error events

Example log record:

{
  "timestamp": "2025-10-18T13:25:41Z",
  "service": "lexia-agent-core",
  "level": "info",
  "message": "Agent responded successfully",
  "duration_ms": 1650
}

Logs are streamed to Lexia's centralized collector and can be queried via:

lexia logs tail --service agent-core

Image: Dashboard visualization of logs and metrics aggregated by agent and environment

3. Alerts & Thresholds

Set performance or safety thresholds to trigger real-time alerts.

Example configuration:

alerts:
  - name: "High Error Rate"
    condition: "error_rate > 5%"
    action: "email"
  - name: "Slow Response"
    condition: "avg_latency > 3000ms"
    action: "slack"

Alerts can be routed to:

Email
Slack / Teams
PagerDuty
Better Stack Alerts

Image: Flow diagram showing alert triggers from Lexia → Slack + Email notifications

4. System Health Dashboard

The Lexia Dashboard → Safety → Observability section shows:

Uptime percentage
API and DB latency
CPU and memory usage
Guardrail and error incident graphs

Each metric links back to corresponding Tracing data for root cause analysis.

Image: Screenshot mockup of Lexia Observability dashboard with charts and alerts summary

5. Integrations (Better Stack, Wazuh)

You can extend observability through external providers.

a. Better Stack

Real-time log streaming
Custom dashboards
Uptime monitoring

Integration:

lexia observability connect betterstack --api-key <API_KEY>

b. Wazuh

Security event correlation
File integrity monitoring
Threat detection