FrankChen021 commented on code in PR #19392:
URL: https://github.com/apache/druid/pull/19392#discussion_r3168197620
##########
extensions-contrib/statsd-emitter/src/main/resources/defaultMetricDimensions.json:
##########
@@ -101,19 +101,33 @@
"segment/moved/bytes" : { "dimensions" : ["dataSource", "taskType"], "type"
: "count" },
"segment/nuked/bytes" : { "dimensions" : ["dataSource", "taskType"], "type"
: "count" },
- "segment/assigned/count" : { "dimensions" : ["tier"], "type" : "count" },
- "segment/moved/count" : { "dimensions" : ["tier"], "type" : "count" },
- "segment/dropped/count" : { "dimensions" : ["tier"], "type" : "count" },
- "segment/deleted/count" : { "dimensions" : ["tier"], "type" : "count" },
- "segment/unneeded/count" : { "dimensions" : ["tier"], "type" : "count" },
+ "segment/assigned/count" : { "dimensions" : ["dataSource", "tier"], "type" :
"count" },
+ "segment/moved/count" : { "dimensions" : ["dataSource", "tier"], "type" :
"count" },
+ "segment/dropped/count" : { "dimensions" : ["dataSource", "tier"], "type" :
"count" },
+ "segment/deleted/count" : { "dimensions" : ["dataSource"], "type" : "count"
},
+ "segment/unneeded/count" : { "dimensions" : ["dataSource", "tier"], "type" :
"count" },
"segment/unavailable/count" : { "dimensions" : ["dataSource"], "type" :
"gauge" },
"segment/underReplicated/count" : { "dimensions" : ["dataSource", "tier"],
"type" : "gauge" },
"segment/loadQueue/size" : { "dimensions" : ["server"], "type" : "gauge" },
- "segment/loadQueue/failed" : { "dimensions" : ["server"], "type" : "gauge" },
"segment/loadQueue/count" : { "dimensions" : ["server"], "type" : "gauge" },
"segment/dropQueue/count" : { "dimensions" : ["server"], "type" : "gauge" },
"segment/size" : { "dimensions" : ["dataSource"], "type" : "gauge" },
"segment/overShadowed/count" : { "dimensions" : [], "type" : "gauge" },
+ "segment/assignSkipped/count" : { "dimensions" : ["dataSource", "server",
"tier", "description"], "type" : "count" },
+ "segment/dropSkipped/count" : { "dimensions" : ["dataSource", "server",
"tier", "description"], "type" : "count" },
+ "segment/moveSkipped/count" : { "dimensions" : ["dataSource", "server",
"tier", "description"], "type" : "count" },
+
+ "segment/count" : { "dimensions" : ["dataSource"], "type" : "gauge" },
+ "segment/availableDeepStorageOnly/count" : { "dimensions" : ["dataSource"],
"type" : "gauge" },
+ "segment/unneededEternityTombstone/count" : { "dimensions" : ["dataSource"],
"type" : "gauge" },
+ "segment/clone/assigned/count" : { "dimensions" : ["dataSource"], "type" :
"count" },
+ "segment/clone/dropped/count" : { "dimensions" : ["dataSource"], "type" :
"count" },
Review Comment:
[P2] Clone-drop metric drops the emitted server dimension
`segment/clone/dropped/count` is emitted from
`CloneHistoricals.dropSegmentFromTargetServer` with only the `server`
dimension, but this new StatsD mapping asks for `dataSource`.
`DimensionConverter` only forwards dimensions present on the event, so
StatsD/DogStatsD will emit this newly mapped metric without either dimension
and aggregate clone drops across all target servers. Map `server`, or emit
`dataSource` as well if datasource-level reporting is intended.
##########
extensions-contrib/statsd-emitter/src/main/resources/defaultMetricDimensions.json:
##########
@@ -178,10 +192,18 @@
"tier/replication/factor" : { "dimensions" : ["tier"], "type" : "gauge" },
"tier/historical/count" : { "dimensions" : ["tier"], "type" : "count" },
+ "tier/storage/capacity" : { "dimensions" : ["tier"], "type" : "gauge" },
+ "tier/historical/clone/count" : { "dimensions" : ["tier"], "type" : "count"
},
+
"compact/task/count" : { "dimensions" : [], "type" : "count" },
"compactTask/maxSlot/count" : { "dimensions" : [], "type" : "count" },
"compactTask/availableSlot/count" : { "dimensions" : [], "type" : "count" },
+ "compact/createJobs/time" : { "dimensions" : [], "type" : "timer" },
+ "compact/createJobs/count" : { "dimensions" : [], "type" : "count" },
+ "compact/runScheduler/time" : { "dimensions" : [], "type" : "timer" },
+ "compactTask/cancelled/count" : { "dimensions" : [], "type" : "count" },
Review Comment:
[P2] Cancelled compaction tasks lose datasource labels
`CompactSegments` records `compactTask/cancelled/count` with
`RowKey.of(Dimension.DATASOURCE, dataSource)`, but the added StatsD mapping has
an empty dimension list. This silently strips datasource from the emitted
StatsD/DogStatsD metric, so cancellations for different datasources are
collapsed into one series. Include `dataSource` in the mapping.
##########
extensions-contrib/statsd-emitter/src/main/resources/defaultMetricDimensions.json:
##########
@@ -199,7 +221,22 @@
"service/heartbeat" : { "dimensions" : ["leader", "workerVersion",
"category", "status", "taskId", "groupId", "dataSource", "taskStatus" ], "type"
: "count" },
+ "metadata/kill/compaction/count" : { "dimensions" : [], "type" : "count" },
+ "metadata/kill/supervisor/count" : { "dimensions" : [], "type" : "count" },
+ "metadata/kill/rule/count" : { "dimensions" : [], "type" : "count" },
+ "metadata/kill/segmentSchema/count" : { "dimensions" : [], "type" : "count"
},
+ "metadata/kill/audit/count" : { "dimensions" : [], "type" : "count" },
+ "metadata/kill/datasource/count" : { "dimensions" : [], "type" : "count" },
+ "kill/eligibleUnusedSegments/count" : { "dimensions" : ["dataSource"],
"type" : "gauge" },
+ "kill/pendingSegments/count" : { "dimensions" : ["dataSource"], "type" :
"count" },
+
"killTask/availableSlot/count" : { "dimensions" : [], "type" : "count" },
"killTask/maxSlot/count" : { "dimensions" : [], "type" : "count" },
- "killTask/task/count" : { "dimensions" : [], "type" : "count" }
+ "kill/task/count" : { "dimensions" : [], "type" : "count" },
+
+ "segment/balancer/compute/error" : { "dimensions" : ["tier", "dataSource",
"description"], "type" : "count" },
+
+ "config/brokerSync/time" : { "dimensions" : [], "type" : "timer" },
Review Comment:
[P2] Broker sync metrics collapse per-broker failures
`BaseDynamicConfigSyncer` emits `config/brokerSync/time` and
`config/brokerSync/error` with `configType` and `server`, and
`config/brokerSync/total/time` with `configType`, but the new mappings for
these metrics use no dimensions. StatsD will therefore aggregate sync latency
and errors across config types and brokers, hiding which broker/config is
failing. Add the emitted dimensions to the mappings.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]