GTFS-RT missing 3 and 6 trains today

[Bo]

Hi! Seems like the 3 and 6 trains have been missing realtime arrivals since midnight? Curious if this was planned or unplanned; I don't see anything in the mta alerts that would suggest realtime arrivals are unavailable today, and trip_replacement_period continues to advance for both lines.

Thanks!

[Dima Kishenko]

We providing school buses 3 6 school buses 

On Thu, Sep 17, 2026, 10:49 AM Bo <b...@lord.io> wrote:

Hi! Seems like the 3 and 6 trains have been missing realtime arrivals since midnight? Curious if this was planned or unplanned; I don't see anything in the mta alerts that would suggest realtime arrivals are unavailable today, and trip_replacement_period continues to advance for both lines.

Thanks!

--
You received this message because you are subscribed to the Google Groups "mtadeveloperresources" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mtadeveloperreso...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mtadeveloperresources/62827576-28e0-4217-8f06-d78be65dd724n%40googlegroups.com.

[Bo Lord]

That's a lot of school buses to make up for 6 train weekday ridership...fortunately it seems like the feed for both the 3 and 6 is back! Thanks.

>> To view this discussion visit https://groups.google.com/d/msgid/mtadeveloperresources/62827576-28e0-4217-8f06-d78be65dd724n%40googlegroups.com <https://groups.google.com/d/msgid/mtadeveloperresources/62827576-28e0-4217-8f06-d78be65dd724n%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> You received this message because you are subscribed to the Google
> Groups "mtadeveloperresources" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to mtadeveloperreso...@googlegroups.com.
> To view this discussion visit

> https://groups.google.com/d/msgid/mtadeveloperresources/CAFu3EUqoQvTU_CBLSZ_FF%3DNjD1HJV9y6mkFsdkGZ1D%2BntPD1wg%40mail.gmail.com
> <https://groups.google.com/d/msgid/mtadeveloperresources/CAFu3EUqoQvTU_CBLSZ_FF%3DNjD1HJV9y6mkFsdkGZ1D%2BntPD1wg%40mail.gmail.com?utm_medium=email&utm_source=footer>.

[Nick Spreen]

We saw this too, after a user reported missing arrivals at 116 St on Lexington Avenue. I inspected the raw feed and found that the problem extended across the 6 line. Does anyone at MTA know what caused it, and whether it was planned maintenance or an unplanned data outage?

Here are the details in case they help trace it. All times below are EDT on September 17, 2026.

Feed endpoint:
https://api-endpoint.mta.info/Dataservice/mtagtfsfeeds/nyct%2Fgtfs

Morning snapshot, header timestamp 09:48:35 EDT / 13:48:35 UTC, Unix 1789652915, GTFS-RT version 1.0:

The feed contained only three TripUpdates with route_id 6, all southbound, and none had future stop predictions. At 116 St, stop_id 622S, their arrival and departure times were identical and already more than eight hours old:

- 000700_6..S01R: 00:42:49 EDT
- 002200_6..S01R: 00:53:42 EDT
- 003950_6..S01R: 01:09:19 EDT

All three used start_date 20260917. Neither their TripDescriptor nor their 622S StopTimeUpdate explicitly set schedule_relationship; there was no explicit SKIPPED indication for that stop. There were no 622N updates. The same snapshot contained three route_id 3 TripUpdates, also with no future stop predictions. Other routes, including 1, 2, 4, 5 and 7, still had future predictions.

Counts of future stop updates near 116 St, northbound / southbound:

- 125 St, 621: 27 / 17
- 116 St, 622: 0 / 0
- 110 St, 623: 0 / 0
- 103 St, 624: 0 / 0
- 96 St, 625: 0 / 0
- 86 St, 626: 24 / 17

The express stops still had 4/5 predictions, which initially made this look like a problem with an individual local stop. The feed header was current despite the stale 6 data.

Around the same time, mta.info showed no active alerts for the 6, but its 116 St arrival panel said, "Information is not available for the 6 line at this time." AP Transit and NYC Subway Status also showed no arrivals for that station. Citymapper showed departures in both directions, although I could not verify whether those were live predictions or its documented scheduled-time fallback.

On our next check at 22:13:26 EDT, the feed header was 22:13:22, about five seconds old. There were 26 route_id 6 trips and 16 route_id 3 trips with future predictions. 116 St had nine northbound and seven southbound 6 predictions, and the nearby local stops had predictions again. This was only a later spot check, not the recovery time; I see Bo had already reported recovery earlier in this thread. We did not monitor continuously and cannot establish the exact outage duration.

Our app currently relies on live predictions and hides stations once their arrivals are empty, so the missing data also made the station disappear for a user. We are looking at keeping stations visible with an explicit data-unavailable state, and possibly adding clearly labeled scheduled departures.

Could MTA share the cause and actual outage window? Is there a feed-health notice or route-level signal consumers should use when predictions disappear but the feed header remains current? Bo's observation that trip_replacement_period kept advancing makes this especially hard to distinguish from an intentional service change. Guidance on the intended fallback would help us handle the next occurrence.

Thanks,
Nick