faker

Faker Service - Armis API Emulator + BGP/BMP Simulator

The faker service emulates the Armis API for testing and reproducing pipeline issues with large datasets. It can also simulate FRR-like BGP activity and export BMP to an Arancini collector for end-to-end ingest testing.

Features

• Generates 25,000 fake devices with unique IPs and metadata
• Emulates Armis API endpoints (/api/v1/access_token/ and /api/v1/search/)
• Each device has:
  • Unique IP address from different ranges (10.0.x.x, 10.1.x.x, 172.16.x.x, 172.17.x.x, 192.168.x.x)
  • Variable number of MAC addresses (1-50, with most having 1-10)
  • Armis device metadata (ID, name, type, OS, manufacturer, etc.)
• Supports pagination for large dataset testing
• Consistent responses for reproducible testing

Running Locally

# Build
go build -o faker ./cmd/faker

# Run
./faker
# Server starts on :8080

Running with Docker

# Build the image
docker build -f cmd/faker/Dockerfile -t serviceradar-faker .

# Run the container
docker run -p 8080:8080 serviceradar-faker

Running with Docker Compose

The faker service is defined in docker-compose.dev.yml (dev-only stack):

# Start faker with the dev compose file
docker compose -f docker-compose.dev.yml up faker

# Or start all dev services
docker compose -f docker-compose.dev.yml up

API Endpoints

Get Access Token

curl -X POST http://localhost:8080/api/v1/access_token/

Search Devices

# Get first 100 devices
curl "http://localhost:8080/api/v1/search/?aql=in:devices&length=100&from=0"

# Get next page
curl "http://localhost:8080/api/v1/search/?aql=in:devices&length=100&from=100"

BGP/BMP Simulation (Arancini Path)

This mode is disabled by default and is intended for demo/test environments.

Prerequisites:

• gobgpd and gobgp must be in PATH.
• Arancini BMP collector must be reachable at simulation.bgp.bmp_collector_address.

Enable by setting simulation.bgp.enabled=true in cmd/faker/config.json (or your deployed config).

Key config fields:

• simulation.bgp.bmp_collector_address (example: 127.0.0.1:11019)
• simulation.bgp.manage_daemon (true: faker starts/stops gobgpd, false: faker controls external gobgpd)
• simulation.bgp.gobgp_api_address (example: serviceradar-gobgp:50051 for external daemon control)
• simulation.bgp.local_asn and simulation.bgp.router_id
• simulation.bgp.peers (FRR-like defaults are included)
• simulation.bgp.advertised_prefixes (defaults: 23.138.124.0/24, 2602:f678::/48)
• simulation.bgp.publish_interval, outage_interval, outage_duration_min, outage_duration_max

Smoke test flow:

1. Start Arancini collector with BMP listen enabled.
2. Start faker with BGP simulation enabled.
3. Confirm faker logs show route announce/withdraw events and periodic outage windows.
4. Confirm Arancini collector logs show incoming BMP session/messages from faker's GoBGP daemon.

Using with Embedded Sync (Agent)

Configure an Armis integration source to point to the faker service:

{
  "sources": {
    "faker-test": {
      "type": "armis",
      "endpoint": "http://faker:8080",
      "partition": "test",
      "credentials": {
        "username": "test",
        "password": "test"
      },
      "queries": [
        {
          "label": "all-devices",
          "query": "in:devices",
          "sweep_modes": ["icmp", "tcp"]
        }
      ],
      "page_size": 1000
    }
  }
}

Testing Large Datasets

The faker service is designed to help reproduce issues with large datasets:

1. Memory Usage: With 25,000 devices, the service uses approximately 20-30MB of memory
2. Pagination: Test pagination with different page sizes (100, 500, 1000)
3. MAC Address Arrays: Devices have varying numbers of MAC addresses to test array handling
4. Consistent Data: Device data is generated deterministically at startup for consistent testing

Debugging Pipeline Issues

To reproduce the sweep.json malformation issue:

1. Start the faker service
2. Configure the embedded sync runtime (via Integrations UI) to use faker as the Armis endpoint
3. Ensure the agent is running so it can fetch all 25,000 devices
4. Monitor KV store writes for malformed sweep.json
5. Check agent's ability to read the sweep configuration

Performance Characteristics

• Startup: ~1 second to generate 25,000 devices
• Memory: ~30MB for 25,000 devices in memory
• Response time: < 10ms for paginated queries
• Max page size: 1000 devices per request