At a glance
Top 10 Slowest Queries (digest) ranks the ten worst-performing query patterns on your MariaDB instance over the last 24 hours, grouped by normalised digest (the query shape with literals stripped). It is the single most actionable performance card for a DBA: it tells you exactly which statements to optimise first. Each row carries the digest text, total execution time, average latency, and call count, so you can tell the difference between a slow query run once and a fast query run a million times.
| What it tracks | The ten query digests with the highest aggregate cost on the selected instance, broken down by row. A digest is the normalised query shape, so SELECT * FROM orders WHERE id = 42 and ... id = 99 collapse into one entry. |
| Data source | performance_schema.events_statements_summary_by_digest, optionally enriched by the slow query log when slow_query_log = ON and long_query_time is set. |
| Time window | 24h rolling. |
| Alert trigger | None. This card is diagnostic, not alerting; pair it with Slow-Query Rate % for the threshold view. |
| Roles | engineering, operations |
What it tracks
The card reads fromperformance_schema.events_statements_summary_by_digest, the Performance Schema table MariaDB maintains in memory for every distinct query shape it has seen since the last counter reset. For each digest it surfaces SUM_TIMER_WAIT (total time spent), COUNT_STAR (how many times the pattern ran), AVG_TIMER_WAIT (mean latency), and SUM_ROWS_EXAMINED (the scan cost). Vortex IQ ranks by total time spent over the trailing 24 hours, because that is the metric that best correlates with where tuning effort pays off: a query that takes 80ms but runs 400,000 times a day costs the engine far more wall-clock than a 12-second report that runs twice. Use the digest text to locate the statement in your ORM or stored procedures, then reach for EXPLAIN and index analysis. A row that scans millions of rows to return a handful is almost always a missing or unused index.
Reconciling against the source
To verify against MariaDB’s own tooling, runSELECT DIGEST_TEXT, COUNT_STAR, SUM_TIMER_WAIT/1000000000000 AS total_sec, AVG_TIMER_WAIT/1000000000 AS avg_ms FROM performance_schema.events_statements_summary_by_digest ORDER BY SUM_TIMER_WAIT DESC LIMIT 10;. On managed services use the equivalent: AWS RDS / SkySQL Performance Insights, or pt-query-digest against the slow log. Note that Performance Schema counters reset on restart, so a freshly restarted node shows a shorter window than Vortex IQ’s full 24h.