Probe Instructions

What are Probes?

Probes are small software agents that you install on your own infrastructure or remote servers. They perform network tests and send the results in real-time to this dashboard.

Total Visibility

Get the exact view of what is happening in different points of presence (PoPs) or in your end client's network.

Agile Troubleshooting

Identify if high latency is a local, IP transit or destination issue by testing from multiple origins.

Implementation Guide

Prepare the Environment

Create a dedicated directory for the probe on your Linux server with Docker installed.

$ mkdir tracelog-probe

$ cd tracelog-probe

Configure Docker Compose

Create the docker-compose.yml file with the correct environment variables.

version: '3.8'

services:
  probe:
    image: evocodeia/tracelog-probe:latest
    container_name: tracelog-probe
    restart: always
    network_mode: "host"
    dns:
      - 8.8.8.8       # Google
      - 1.1.1.1       # Cloudflare
      - 8.8.4.4       # Google Backup
      - 1.0.0.1       # Cloudflare Backup
      - 9.9.9.9       # Quad9
      - 208.67.222.222 # OpenDNS
    environment:
      - API_BASE_URL=https://tracelog.evocode.ia.br/api
      - SENSOR_KEY=SUA_CHAVE_DE_API
      - CHECK_INTERVAL_MINUTES=5
      - MAX_WORKERS=5 # Opcional: Testes simultâneos (Default: 5)
      - MTR_INTERVAL_SECONDS=0.5 # Opcional: Intervalo entre pacotes (Default: 0.5s)

Attention

The SENSOR_KEY must be generated in the "Network > Probes" menu of this dashboard. Without this key, the probe will not be able to authenticate with our servers.

Tip: You can now override the MTR Packet Interval for specific targets directly in the Target settings. The MTR_INTERVAL_SECONDS variable serves as a global fallback.

Start and Monitor

Start the container and follow the logs in real-time to validate communication.

$docker logs -f tracelog-probe

Agente NOC Compatibility

Remote probe deployment continues to happen through Docker. For teams that already use Agente NOC, TraceLog can expose telemetry and statistics inside the agent as an optional compatibility layer.

Operational continuity

TraceLog remains the observability platform while Agente NOC becomes an extra interface for consulting route changes, latency, and evidence.

Data access with context

Customers who already operate with Agente NOC can query TraceLog measurements without duplicating dashboards or changing probe deployment.

Need the full compatibility overview?

See compatibility page

Performance Tuning

Why is the interval configured here?

TraceLog delegates execution to distributed probes. The check interval is defined globally on the probe (CHECK_INTERVAL) to ensure consistency and respect processing capacity (Workers). Defining individual intervals per target would cause fragmentation and capacity calculation issues.

We recommend always configuring the probe interval with a safety margin (e.g., +20%) above the calculated minimum.

By default, the probe processes 5 targets simultaneously. If you are monitoring a large number of targets (e.g., > 50) with a single probe, you may need to adjust the MAX_WORKERS to ensure all checks complete within the configured interval.

Optimization Formula

Scenario

Targets100
Interval5 min (300s)
Avg MTR Time~15s

Recommendation

MAX_WORKERS = (100 * 15s) / 300s ≈ 5

In this case, 5 workers are sufficient. For 500 targets, you would increase to 25 workers.

Networking Tip

The NET_RAW and NET_ADMIN flags are fundamental for MTR and Traceroute to run in privileged mode in the kernel.

Auto-Update

Keep your probes up to date by running "docker-compose pull" periodically to load new diagnostic features.

Versioning

TraceLog now automatically tracks your probe version. Check in "Network > Probes" if you are running the latest stable version of the telemetry engine.Current probe version: vdev

Back to Center