textfile-exporter

A lightweight HTTP server that reads metrics from a text file and exposes them in Prometheus exposition format.

Useful for exporting deployment metadata, environment variables, or any static/slowly-changing values to Prometheus without writing a full exporter.

How it works

1. You write metrics to a file in Prometheus text format (e.g. from a deploy script or CI/CD pipeline)
2. textfile-exporter serves that file at /metrics
3. Prometheus scrapes the endpoint as usual

Quick start

# Write metrics to a file
cat > metrics.txt << EOF
# HELP deploy_info Deployment metadata
# TYPE deploy_info gauge
deploy_info{version="1.2.3",env="prod",commit="abc123"} 1

# HELP deploy_timestamp Unix timestamp of last deployment
# TYPE deploy_timestamp gauge
deploy_timestamp 1713436800
EOF

# Run the exporter
./textfile-exporter
# or
docker run -v $(pwd)/metrics.txt:/metrics.txt -p 8080:8080 textfile-exporter

$ curl localhost:8080/metrics
# HELP deploy_info Deployment metadata
# TYPE deploy_info gauge
deploy_info{version="1.2.3",env="prod",commit="abc123"} 1
...

Configuration

All settings are configured via environment variables:

Variable	Default	Description
SERVE_ADDR	:8080	Address and port to listen on
METRICS_FILE_PATH	metrics.txt	Path to the metrics file
METRICS_ENDPOINT	/metrics	HTTP path to expose metrics at
REFRESH_INTERVAL	0	How often to reload the file (0 = load once at startup). Accepts Go duration strings: 30s, 1m, 5m
LOG_OUTPUT	stdout	Log destination (stdout or stderr)

Docker

docker build -t textfile-exporter .

# Mount your metrics file and run
docker run \
  -v /path/to/metrics.txt:/metrics.txt \
  -e METRICS_FILE_PATH=/metrics.txt \
  -e REFRESH_INTERVAL=30s \
  -p 8080:8080 \
  textfile-exporter

Use case: deploy-time metrics

Generate metrics.txt as part of your deployment pipeline:

#!/bin/bash
cat > /var/metrics/deploy.txt << EOF
# HELP deploy_info Current deployment info
# TYPE deploy_info gauge
deploy_info{version="${APP_VERSION}",env="${ENVIRONMENT}",commit="${GIT_COMMIT}"} 1

# HELP deploy_timestamp Unix timestamp of last successful deployment
# TYPE deploy_timestamp gauge
deploy_timestamp $(date +%s)

# HELP build_number CI build number
# TYPE build_number gauge
build_number ${BUILD_NUMBER}
EOF

Run textfile-exporter with REFRESH_INTERVAL=30s so it picks up the new file after each deploy without restarting.

Prometheus config

scrape_configs:
  - job_name: deploy_metrics
    static_configs:
      - targets: ['localhost:8080']

Metrics file format

Standard Prometheus text format:

# HELP metric_name Description of the metric
# TYPE metric_name gauge
metric_name{label="value"} 42
metric_name_with_timestamp{label="value"} 42 1713436800000

Supported types: gauge, counter, untyped.

Endpoints

Path	Description
/metrics	Prometheus metrics (configurable via METRICS_ENDPOINT)
/	Health check — returns 200 OK