Hi,
After upgrading from Metabase 0.48.4 → 0.55.8, several of our internal monitoring queries stopped returning data.
We rely on these queries to track things like query usage/performance and alert/pulse delivery health.
We checked the 0.49–0.55 release notes and changelogs but couldn’t locate any documented internal-schema changes that would explain the issue, maybe we’re missing something.
Example query now returning empty results
We have a query that used to report pulse/alert failures and successes by joining internal tables (query_execution, pulse, report_card)
SELECT
started_at::date AS "Execution Date",
CASE WHEN error IS NULL THEN 'Success' ELSE 'Failed' END AS "Execution Status",
COUNT(*)
FROM query_execution
INNER JOIN pulse ON query_execution.pulse_id = pulse.id
INNER JOIN report_card ON report_card.id = card_id
INNER JOIN metabase_database ON metabase_database.id = query_execution.database_id
WHERE TRUE
[[AND {{started_at}}]]
[[AND CASE WHEN pulse.name IS NULL THEN 'Alert' ELSE 'Pulse' END ILIKE {{type}}]]
[[AND CASE WHEN error IS NULL THEN 'Success' ELSE 'Failed' END ILIKE {{status}}]]
[[AND {{database}}]]
[[AND {{card}}]]
[[AND {{card_id}}]]
[[AND {{pulse}}]]
[[AND (running_time/1000) >= {{min_runtime}}]]
GROUP BY 1, 2
ORDER BY 1, 2 DESC;
Before the upgrade, this produced daily counts of successful and failed Pulse/Alert executions.
After the upgrade, no results.
Thanks,