{
  "detail": "The date(s) you used are valid, but we either do not have data for those date(s), or the user_central_id you asked for does not have data for those dates. Please check documentation for more information",
  "method": "get",
  "status": 404,
  "title": "Not Found",
  "type": "about:blank",
  "uri": "/metrics/pageviews/v3/top_pages_per_editor/12345/monthly/20251201/20260501"
}

Contrast this to https://wikimedia.org/api/rest_v1/metrics/pageviews/v3/top_pages_per_editor/12345/monthly/20251101/20260501 which returns data for 2025-11. Months 2025-02 to 2025-11 return data, but nothing after.

Everything looks fine in Airflow. However, examining the Cassandra loading Spark job logs shows these errors:

sudo -u analytics      yarn logs -appOwner analytics -applicationId application_1773845446826_818239
[...]
26/05/01 02:29:28 INFO Native: Unable to load JNR native implementation. This could be normal if JNR is excluded from the classpath
java.lang.NoClassDefFoundError: jnr/posix/POSIXHandler
	at com.datastax.oss.driver.internal.core.os.Native$LibcLoader.load(Native.java:42)
[...]
Caused by: java.lang.ClassNotFoundException: jnr.posix.POSIXHandler
	at java.net.URLClassLoader.findClass(URLClassLoader.java:387)
[...]
26/05/01 02:39:03 INFO Executor: Finished task 2.0 in stage 0.0 (TID 4). 1786 bytes result sent to driver
26/05/01 02:40:07 ERROR CoarseGrainedExecutorBackend: RECEIVED SIGNAL TERM

I can't tell if that error is benign or actually something wrong.

Related Objects

Mentioned In: T426373: Mediaviews Analysis returns API not found error
T414020: Impact Module: Investigate using Wikimedia Analytics API
Mentioned Here: T414020: Impact Module: Investigate using Wikimedia Analytics API

Event Timeline

Ottomata created this task.May 15 2026, 3:30 PM

Restricted Application added a subscriber: • Aklapper. · View Herald TranscriptMay 15 2026, 3:30 PM

Ottomata updated the task description. (Show Details)May 15 2026, 3:32 PM

I gave up my root access a few months ago, so I can't query cassandra directly anymore to see if the data is there. I can't find any page-analytics AQS service logs anywhere (neither in logstash nor via k8s logs).

We need @Eevans or someone with cassandra access to debug further.

Eric, the (Hive referenced) AQS cassandra table name is aqs.analytics.pageviews_top_pages_per_editor. The query that loads data looks something like:

INSERT INTO aqs.analytics.pageviews_top_pages_per_editor
SELECT 
    user_central_id,
    granularity,
    wiki_id,
    page_id,
    rank,
    top_k,
    view_count,
    dt
FROM
    wmf_readership.pageviews_top_pages_per_editor
WHERE
    dt = CAST(TO_DATE('2026-04-01', 'yyyy-MM-dd') AS TIMESTAMP)
    AND granularity = 'monthly'
;

Can you tell if there is any data in that table after 2026-11?

Ottomata mentioned this in T414020: Impact Module: Investigate using Wikimedia Analytics API.May 15 2026, 3:38 PM

Oh, Eric commented in Slack. Investigating more...

So, the data is not in Cassandra. That means the cassandra_load_pageview_top_pages_per_editor_monthly.load_cassandra Airflow task is failing silently.

Will need more investigation.

Ahoelzl assigned this task to amastilovic.May 15 2026, 3:46 PM

Ahoelzl triaged this task as High priority.

Ahoelzl updated Other Assignee, added: amastilovic.

Ahoelzl edited projects, added: Data-Engineering (Q4 FS25/26 April 1st - June 30st); removed: Data-Engineering.

Ahoelzl moved this task from Next Up to Urgent on the Data-Engineering (Q4 FS25/26 April 1st - June 30st) board.

Ahoelzl updated Other Assignee, removed: amastilovic.

Looking into this!

MusikAnimal mentioned this in T426373: Mediaviews Analysis returns API not found error.May 15 2026, 5:47 PM

Data is available now in Cassandra / API

% curl -X GET "https://wikimedia.org/api/rest_v1/metrics/pageviews/v3/top_pages_per_editor/12345/monthly/20241201/20260601" -H "accept: application/json" | jq '.items[].timestamp'
"2025-02-01T00:00:00.000Z"
"2025-03-01T00:00:00.000Z"
"2025-04-01T00:00:00.000Z"
"2025-05-01T00:00:00.000Z"
"2025-06-01T00:00:00.000Z"
"2025-07-01T00:00:00.000Z"
"2025-08-01T00:00:00.000Z"
"2025-09-01T00:00:00.000Z"
"2025-10-01T00:00:00.000Z"
"2025-11-01T00:00:00.000Z"
"2025-12-01T00:00:00.000Z"
"2026-01-01T00:00:00.000Z"
"2026-02-01T00:00:00.000Z"
"2026-03-01T00:00:00.000Z"
"2026-04-01T00:00:00.000Z"

Michael subscribed.May 19 2026, 11:11 AM

Ahoelzl reassigned this task from amastilovic to mforns.May 26 2026, 3:48 PM

mforns moved this task from Urgent to Done on the Data-Engineering (Q4 FS25/26 April 1st - June 30st) board.Jun 1 2026, 3:13 PM

Ahoelzl closed this task as Resolved.Tue, Jun 30, 4:28 PM

AQS pageviews/v3/top_pages_per_editor not returning dataClosed, ResolvedPublicActions

Description

Related Objects

Event Timeline

AQS pageviews/v3/top_pages_per_editor not returning data
Closed, ResolvedPublic
Actions