VDL2 ingest down for a cluster of EU feeders since 2026-06-06 09:54 UTC (~3x global rate drop)

[salsburg]

Flagging an observation from the data side — I feed airframes (JS-KPBI) and pull the /messages firehose for coverage monitoring. Everything below is reproducible from your own /messages and /stations endpoints.

Summary: At ~09:54 UTC on 2026-06-06 the global message rate stepped down from ~70/s to ~13/s and has held there since (29h+ as of 2026-06-07 14:37 UTC). The loss is concentrated in VDL2 from European feeders. US VDL2, ACARS, HFDL, and Iridium are unaffected (some up).

A cluster of EU VDL2 feeders stopped reporting inside the same ~3-minute window, and /stations still lists them active:

station	site	lastReportAt
BB-EDDN1-VDL2	Munich	2026-06-06T09:54:58.230Z
AM-LSZH-VDL2	Zurich	2026-06-06T09:54:01.379Z
KAR-LEVC-VDLM	Valencia	2026-06-06T09:53:56.986Z
SS-EGBB-VDL-M2	Birmingham	2026-06-06T09:54:08.524Z
PYC-LFPL-VDLM2	Paris	2026-06-06T09:53:59.298Z
MCC01-EGJJ-VDLM2	Jersey	2026-06-06T09:56:06.142Z

(~16 VDL2 stations currently show lastReportAt frozen in the 09:40–10:30 UTC window on 2026-06-06.)

What points to ingest rather than the feeders:

• Same site, two chains: SS-EGBB-ACARS is still reporting normally (last seen 2026-06-07T11:44:11.529Z), while SS-EGBB-VDL-M2 froze at 09:54. Same building/power/internet, only VDL2 stopped.
• Independent corroboration: a separate public ACARS aggregator (adsb.lol) shows no drop over the same window and still carries these European sites, so the feeders appear to be transmitting fine.

Could be expected on your end, but it looked like a VDL2 ingest issue worth flagging. Happy to share the full per-station list and the hourly timeline.

[salsburg]

@kevinelliott flagging in case this slipped by — still down ~49h on. The 16 EU VDL2 feeders remain frozen at their 09:54 UTC lastReportAt (2026-06-06), and the global /messages rate is holding ~15/s vs ~60/s normal (US / ACARS / HFDL / Iridium all unaffected). Happy to share the full per-station list + hourly timeline if it helps.

[kevinelliott]

Hey @salsburg thanks for reporting. Let me dig in and see what might be going on. Give me a few.

[kevinelliott]

And yes, include any data you have!

[salsburg]

Thanks Kevin! Here's the data, all from your own /messages + /stations so you can verify directly.

16 VDL2 feeders stopped within a ~2-minute window and are still frozen 52h later, yet /stations still lists them active:

station	airport	lastReportAt (UTC, 2026-06-06)
BNE-YENO-VDL2	YBBN	09:48:15
VDL-109-77	EGKA	09:53:47
KAR-LEVC-VDLM	LEVC	09:53:56
portsmouthms-VDLM2	EGHF	09:53:57
VDL-M2	EGBB	09:53:58
PYC-LFPL-VDLM2	LFPL	09:53:59
AM-LSZH-VDL2	LSZH	09:54:01
GG-EGJB-VDLM2	EGJB	09:54:05
SS-EGBB-VDL-M2	EGBB	09:54:08
UN-EGNR-VDLM2	EGNR	09:54:10
M7KNI-EGHI-VDL2	EGHI	09:54:22
BB-EDDN1-VDL2	EDDM	09:54:58
RK-ETNG-VDL2	EHBK	09:55:55
KK-EDVR-VDLM2	EDVR	09:56:03
MCC01-EGJJ-VDLM2	EGJJ	09:56:06
NGN-LMML-VDL2	LMML	09:56:28

All Europe/UK (plus one Brisbane), all dumpvdl2/vdlm2dec stations, every lastReportAt between 09:48 and 09:56 UTC on 2026-06-06.

Same-site tell: at Birmingham, SS-EGBB-ACARS is reporting fine right now (last seen 2026-06-08T13:42Z) while SS-EGBB-VDL-M2 froze at 09:54 on the 6th. Same host, ACARS up, VDL2 dead — so it isn't the feeders.

Protocol mix at the cliff (06-06 08:00 vs 12:00 UTC, distinct msgs): VDL2 104,933 → 33,573 (−68%); satcom aero-acars −41%; plain ACARS −21%; HFDL +24%; Iridium +112%. Only VDL2 fell.

One nuance: gross /messages rate looks near-normal at the daily peak (~85/s right now), so the loss hides in volume — the frozen-station list is the real tell, clearest off-peak (06-06 10:00 UTC dropped ~63→26/s; overnight-UTC hours have run ~15/s vs ~50/s the prior weekend).

These feeders are still live on another aggregator, so they're transmitting fine — looks ingest-side, specific to the EU VDL2 path. Happy to send the raw hourly series or anything else.

[salsburg]

Quick correction on one line above: I cited ~85/s "at peak" from a short live sample, but that was a measurement artifact — the /messages latest-id is choppy over short windows (a 30s re-check showed 0 advance, while the newest 100 msgs spanned 3.6s ≈ 28/s). The reliable full-hour numbers have the global rate at ~13–28/s continuously since the 06-06 cliff — i.e. down ~3–4x at all hours, not just off-peak, so the loss does not hide in volume. Everything else stands (the 16 feeders frozen since 09:5x UTC, same-site SS-EGBB-ACARS still reporting, protocol mix VDL2-only). Sorry for the noise on that figure.