New Rule Proposal: Feed-wide threshold of VehiclePosition records having missing/stale timestamp

[evansiroky]

Summary:

A warning should be raised whenever a feed-wide threshold has been met among all VehiclePosition entities that have a missing or stale timestamp.

Steps to reproduce:

Given a Vehicle Position dataset When the validator has scanned each individual VehiclePosition record and detects that a certain feed-wide threshold of those records have a timestamp that is missing or stale Then the validator should flag the Vehicle Position feed in question for having a feed-wide problem of missing or stale timestamps being set

Expected behavior:

The GTFS-Realtime Best Practices state:

Data within a GTFS Realtime feed should not be older than 90 seconds for Trip Updates and Vehicle Positions and not older than 10 minutes for Service Alerts.

The GTFS Validator should flag an error whenever a certain percent of VehiclePosition records do not have a timestamp value added or one that is added but is far enough in the past that it is stale.

Observed behavior:

Timestamps are currently only checked at the feed-level.

etc

• This will need some definition of what percent of VehiclePosition records would need a missing/stale timestamp in order to be flagged.

[e-lo]

Before implementing, need to

• [ ] decide on threshold for % and/or max # of missing timestamps
• [ ] document it in an update to the best practices