RIPEstats apparently broken

This is what the example query from
RIPEstat docs | Announced Prefixes | Docs is
returning:

{
  "messages": [
    [
      "error",
      "There was a problem handling this request. The error has been
logged and we will look into the cause as soon as possible. We
apologise for any inconvenience."
    ],
    [
      "warning",
      "There was a failure working out what the latest available data
point is. Results may be incorrect!"
    ]
  ],
  "see_also": ,
  "version": "1.2",
  "data_call_name": "announced-prefixes",
  "data_call_status": "supported",
  "cached": false,
  "data": {},
  "query_id": "20240802194014-83da148b-94dc-4cd3-9655-7146cf1fee0a",
  "process_time": 7,
  "server_id": "app114",
  "build_version": "live.2024.7.29.212",
  "status": "error",
  "status_code": 500,
  "time": "2024-08-02T19:40:14.012851"
}

I would assume this was not intentional... are any other users out
there affected by this issue ?

Rubens

I have a custom nagios check, ripestat_monitor_routes.py, that has been
returning HTTP Error 500: Internal Server Error for about 8 hours.

Regards,
Robert

USC Information Sciences Institute <http://www.isi.edu/&gt;
Networking and Cybersecurity Division

RIPE conveniently has a status page for this exact incident https://status.ripe.net/incidents/8jm3b6j26kcf

https://status.ripe.net/ - they know and are working to fix things.

What would be really convenient was to propagate the internal HTTP
error to the external handler. It's returning HTTP 200 with this in
the body:

"status": "error",
  "status_code": 500,

Returning HTTP 500 would have made my script happier, as it would move
ahead without that data. But because it was 200, it tried to parse the
data it was looking for...

Rubens

Hi Rubens,

What would be really convenient was to propagate the internal HTTP
error to the external handler. It's returning HTTP 200 with this in
the body:

"status": "error",
  "status_code": 500,

Returning HTTP 500 would have made my script happier, as it would move
ahead without that data. But because it was 200, it tried to parse the
data it was looking for...

We recently became aware of this issue. This behaviour is not consistent between
the various API calls. We will fix it and ensure that API calls that fail return
a relevant HTTP status code.

About the current incident: The underlying database has recovered. Some data is
not up to date yet but should recover over time. We will provide further updates
on the status page [0] until the issue is resolved.

Kind regards,

Ties de Kock
RIPE NCC

[0] RIPE NCC Status - Multiple RIPEstat backends are unavailable