mocktrics-exporter 0.2.0

A FastAPI-based mock Prometheus exporter with configurable metrics and UI.

mocktrics-exporter

Small, configurable Prometheus exporter for generating test metrics. It exposes:

• a Prometheus metrics endpoint for scraping
• a small FastAPI HTTP API for defining/updating metrics at runtime

Works great for demos, integration tests, and performance scenarios where you need predictable metrics behavior (static, ramp, square, sine, gaussian) with labels.

Note: The Python package is named mocktrics_exporter (underscore in the import path).

Install

• From PyPI:
  • pip install mocktrics-exporter
• From source (editable):
  • pip install -e .

Quick Start

• Run the exporter:
  • mocktrics-exporter -f config.yaml
• Alternatively from source:
  • python -m mocktrics_exporter.main -f config.yaml
• Defaults:
  • Metrics served on :8000
  • HTTP API served on :8080

Example Prometheus scrape config:

• job_name: mocktrics
• static_configs: ['localhost:8000']

CLI Options

• -f, --config-file path to YAML configuration file
• -a, --api-port API port (default 8080)
• -m, --metrics-port Prometheus metrics port (default 8000)

Options can also be provided via environment or process managers as needed.

Configuration

• File: config.yaml (optional). If provided, metrics are preloaded at startup.
• Schema overview: each metric defines a name, documentation, optional unit, a list of labels, and a list of values. Values use a kind discriminator and fields specific to the chosen kind (see Supported Value Models below).

Example:

disable_units: true

metrics:
  - name: combined
    documentation: Mock metric for all types
    labels: [type]
    values:
      - kind: static
        value: 100
        labels: [static]
      - kind: ramp
        period: 2m
        peak: 100
        invert: false
        labels: [ramp]
      - kind: square
        period: 2m
        magnitude: 100
        duty_cycle: 50
        labels: [square]
      - kind: sine
        period: 2m
        amplitude: 50
        offset: 50
        labels: [sine]

Supported Value Models

• static: constant numeric value
• ramp: linear ramp up to peak over period, optional invert, offset
• square: square wave with period, magnitude, duty_cycle (0–100), optional invert, offset
• sine: sine wave with period, amplitude, optional offset
• gaussian: random gaussian with mean, sigma

Helpers:

• Duration strings: 1s, 2m, 3h, 1d
• Size strings: 2u, 2m, 2k, 2M, 2G

HTTP API

Base URL is the API port (default http://localhost:8080). Interactive API docs are available at http://localhost:8080/docs (Swagger UI).

Example requests:

# Create a metric
curl -X POST localhost:8080/metric \
  -H 'content-type: application/json' \
  -d '{
    "name": "http_requests",
    "documentation": "Demo metric",
    "unit": "",
    "labels": ["method"],
    "values": [{"kind": "static", "labels": ["GET"], "value": 42}]
  }'

# Add a value
curl -X POST localhost:8080/metric/http_requests/value \
  -H 'content-type: application/json' \
  -d '{"kind": "ramp", "labels": ["POST"], "period": "2m", "peak": 100}'

# List metrics
curl localhost:8080/metric/all

# Delete a value
curl -X DELETE 'localhost:8080/metric/http_requests/value?labels=GET'

# Delete metric
curl -X DELETE localhost:8080/metric/http_requests

Prometheus Metrics

• Metrics endpoint runs on the metrics port (default 8000).
• Each metric is exported as a Gauge with labels as defined.
• Units in the metric name suffix can be disabled with disable_units: true in config.

Development

• Run tests: pytest -q (includes fast API and unit tests)
• Run API with autoreload for local dev: uvicorn mocktrics_exporter.api:api --reload --port 8080
• Code style: Black, isort, autoflake via pre-commit hooks
• Python: >=3.9

License

Apache 2.0. See LICENSE.