trino.execution.time.p99

99th percentile execution time

Dimensions:None

Available on:Native (1)

Interface Metrics (1)

Native

trino.execution.QueryManager.ExecutionTime.FiveMinute.P99

99th percentile query execution time in the last five minutes

Dimensions:None

Sources

trino.execution.QueryManager.ExecutionTime.FiveMinute.P99trino.io

Technical Annotations (26)

Configuration Parameters (6)

retry-policyrecommended: TASK (for large batch queries)

TASK policy best suited for retrying large batch queries with long execution times

-Xmx110Grecommended: 110G

max heap on 125G worker; may need tuning for off-heap memory pressure

-XX:ReservedCodeCacheSizerecommended: 2G

code cache for JIT compiled PageFilter classes; may fill under sustained load

query.max-execution-time

Max query execution time in config.properties; increase if legitimate queries are timing out

query.remote-task.max-error-durationrecommended: 1s

Controls how quickly queries transition to BLOCKED after task failure

query.client.timeoutrecommended: 5m (default)

Does NOT affect the FINISHING state delay despite 5-minute correlation

Error Signatures (1)

QUERY_CANCELEDerror code

CLI Commands (5)

top -H -p <pid>diagnostic

jstack -l <pid>diagnostic

EXPLAIN ANALYZE VERBOSEdiagnostic

SELECT query_id, user, query, create_time, end_time, (end_time - create_time) AS duration_seconds FROM trino_events.trino_queries WHERE end_time IS NOT NULL ORDER BY duration_seconds DESC LIMIT 10;

diagnostic

SELECT user, AVG(end_time - create_time) AS avg_duration FROM trino_events.trino_queries WHERE end_time IS NOT NULL AND create_time >= now() - interval '7' day GROUP BY user ORDER BY avg_duration DESC LIMIT 10;

diagnostic

Technical References (14)

S3 prefixcomponentcardinalityconcepttrino_events.trino_queriescomponentcreate_timecomponentend_timecomponentio.trino.$gen.PageFiltercomponentC2 CompilerThreadcomponentio.trino.operator.ScanFilterAndProjectOperatorcomponentio.trino.parquet.reader.PrimitiveColumnReadercomponentstage 0conceptFINISHINGcomponentBLOCKEDcomponentfault-tolerant executionconcepttrino-python-clientcomponent

Related Insights (9)

Long-running queries have increased failure probability without FTEwarning

▸

Undersized cluster causes slow query performancewarning

▸

Slow query performance from large unpartitioned data scanswarning

▸

Long-running queries hog cluster resourceswarning

▸

Single worker CPU saturation causes cluster-wide query timeoutscritical

▸

Long-running tasks accumulate on single worker causing timeoutscritical

▸

Long-running queries cause resource contentionwarning

▸

Queries canceled when execution time exceeds configured limitwarning

▸

Query hangs in FINISHING state for 5 minutes after fault-tolerant retrywarning

▸