View entry
Tap Metrics monitoring to view the monitoring metrics data of the instance. Use the dropdown menu to select a time period and view detailed metrics data for that period.
Metrics description
SQL_QRS
Metric item | Meaning |
QPS | QPS of SQL queries, including various error queries |
EmptyQps | QPS of normal queries or independent phase one queries on Qrs with empty results |
ErrorQps | Total number of SQL query errors |
GetPlanErrorQps | Number of failures in generating an available iquan plan for SQL queries |
SessionLatency | Total time consumed by SQL queries, including thread pool queuing time and SQL query processing time |
ProcessLatency | Time consumed in processing SQL queries |
Sql plan parsing time | Time consumed in parsing SQL plan |
Number of returned results | Number of returned results |
Length of returned results | Length of returned results |
SQL_Searcher
Metric item | Meaning |
qps | QPS of SQL queries, including various error queries |
Searcher query time | Time consumed in Searcher queries |
QRS
Metric item | Meaning |
qps | Query QPS |
Phase one qps | Phase one query QPS |
Phase two qps | Phase two query QPS |
proc_cpu | CPU utilization of the query node |
proc_mem_rss | Memory usage of the query node |
mem | Memory usage of the query node |
qrsSessionLatencyNormal | Time from session queue export to completion for normal queries/independent phase one/independent phase two on Qrs |
waitLatencyPhase1 | Time spent waiting for asynchronous request results from descendant nodes in phase one on Qrs |
waitLatencyPhase2 | Time spent waiting for asynchronous request results from descendant nodes in phase two on Qrs |
returnCountNormal | Number of normal returns from the query node |
resultLenNormal | Length of normal query results |
emptyQps | QPS of queries with empty results |
syntaxErrorQps | QPS of syntax error requests on Qrs during queries, including errors such as parse clause failure and illegal result format |
RequestPoolWaitTime | Time spent waiting in the worker queue for sessions on the query node |
multiCallLatency | Latency of multiCall invocation on the query node |
early_terminator_qps | QPS of queries that trigger early termination on the query node |
worker_status | Status of the data node (1 indicates normal, 0 indicates abnormal) |
processerrorqps | QPS of process error requests on Qrs during queries, including errors where Processor or ProcessorChain does not exist causing Qrs to be unable to process the request, and Process errors, excluding syntax errors |
Searcher-normal
Metric item | Meaning |
proc_Cpu | CPU utilization of the query node |
proc_Mem_Rss | Memory usage of the query node |
proc_Mem_Use_Ratio | Memory usage of the data node |
disk_use | Disk usage of the data node |
disk_use_ratio | Disk usage ratio of the data node |
patitionDocCount | Number of docs stored in the data node |
partitionMemoryQuotaUse | Memory quota used by the index on the data node |
incIndexMemoryUse | Memory used by incremental index |
oldInMemorySegmentMemoryUse | Memory used by unreleased memory segments |
SegmentCount | Number of segments on the data node |
requestPoolWaitTime | Time spent waiting in the query pool for query requests on the data node |
worker_status | Status of the data node (1 indicates normal, 0 indicates abnormal) |
deletedDocCount | Number of deleted docs on the data node |
Searcher_phase1
Metric item | Meaning |
inQpsPhase1 | QPS in phase one on the data node |
searcherProcessLatencyPhase1 | Phase one query processing latency on the data node |
sessionLatencyPhase1 | Phase one query session latency on the data node |
beforeSearchLatency | Initialization latency of queries on the data node |
rankLatency | Time spent in the first ranking on the data node |
rerankLatency | Time spent in fine ranking on the data node |
matchCount | Number of matched results on the data node, accurate value |
matchDocSize | Actual number of matchDoc returned in phase one on the data node |
seekCount | Total number of seeks in queries on the data node, accurate value |
seekDocCount | Number of seek docs on the data node |
phase1EmptyQps | QPS of empty results in phase one queries on the data node |
phrase1TimeoutQps | QPS of timeout queries in phase one on the data node |
cacheHitRatio | Full cache hit ratio on the data node |
cacheGetNum | QPS of full cache being invoked on the data node |
cachePutNum | QPS of results added to full cache on the data node |
cacheItemNum | Number of items in full cache on the data node |
extraRankCount | Number of docs in the final round of ranking on the data node |
extraRankLatency | Time spent in the final round of ranking on the data node |
aitheta_seek_count | Number of aitheta seeks on the data node |
aitheta_recall_ratio | Recall rate of aitheta on the data node |
aitheta_seek_latency | Latency of aitheta seeks on the data node |
after_search_latency | Time from the end of ranking to the final return of results on the data node |
request_pool_wait_time | Time spent waiting in the query pool for query requests on the data node |
Searcher_phase2
Metric item | Meaning |
inQpsPhase2 | QPS of phase two queries on the data node |
searcherProcessLatencyPhase2 | Phase two query process latency on the data node |
sessionLatencyPhase2 | Phase two query session latency on the data node |
blockCacheHitRatio | Block cache hit ratio on the data node |
fetchSummaryLatency | Time consumed in fetching summary on the data node |
extractSummaryLatency | Time consumed in extracting dynamic summary on the data node |
Searcher_rt_build
Metric item | Meaning |
addQps | QPS of API push add doc in the data node |
deleteQPS | QPS of API push delete doc in the data node |
buildQps | QPS of real-time build index in the data node |
end2EndLatency_API | Latency of API push data |
end2EndLatency_default | Latency metric of API push data |
swiftReadDelay | Swift read latency on the data node |
rtIndexMemoryUse | Real-time memory usage in the data node |
rtIndexMemoryUseRatio | Real-time memory usage ratio in the data node |
dumpSegmentLatency | Latency of dumping real-time data segment during batch switching on the data node |
processedDocSize | Size of processed documents on the data node |