/gtfs-rt-metrics

Primary LanguagePython

gtfs-rt-metrics

This is gtfs-rt-metrics, a project to collect metrics on GTFS-realtime feeds and log them to an InfluxDB time-series database, where they can be visualized with Grafana, Chronograf, or other tools.

Deployment

gtfs-rt-metrics is most easily deployed with Docker and Docker Compose:

  1. Copy config-sample.ini to config.ini and edit as desired. You may want to edit the list of feeds being monitored, change the polling interval, or alter the InfluxDB configuration parameters.
  2. Copy docker-compose-sample.yml to docker-compose.yml. Edit as desired (although no changes are required for the default configuration).
  3. Build the gtfs-rt-metrics image with docker-compose build.
  4. Start gtfs-rt-metrics, InfluxDB, and Grafana with docker-compose up -d.
  5. If required, create the rtmetrics InfluxDB database (launch the influx client and run CREATE DATABASE rtmetrics).

Next Steps

gtfs-rt-metrics will begin collecting the following metrics for each feed into InfluxDB every 20 seconds:

  • count of feed entities
  • count of trip updates, vehicle positions, and alerts
  • count of trip updates per route
  • age of feed header timestamp
  • average age of feed entity timestamp (only applicable to trip updates and vehicle positions)
  • feed response time
  • feed response size

All of the above are stored in an InfluxDB series called feed_fetch, except for the count of trip updates per route, which are stored in the route_fetch series.