Julian Stephen created FLINK-9653:
-
Summary: Add operator name to latency metrics
Key: FLINK-9653
URL: https://issues.apache.org/jira/browse/FLINK-9653
Project: Flink
Issue Type: New Feature
Components: Metrics
Environment: All
Reporter: Julian Stephen
Currently the latency metrics reports latency between subtasks using this
format:
flink_taskmanager_job_latency_source_id_source_subtask_index_operator_id_operator_subtask_index_latency{
host="",instance="",job="",
job_id="",job_name="",operator_id="",operator_subtask_index="0",
quantile="0.99",source_id="",source_subtask_index="0",tm_id=""
The request is to add {{operator_name}} along with {{operator_id}} to the
metric labels.
For a simple job, (e.g., {{source->map->sink) }} you can see two sets of
latency metrics. Each set shows all quantiles like (.5, .95..). Only thing
different between the two sets is the {{operator_id}}. This makes sense
assuming one {{operator_id}} belongs to the {{map}} operator and the other
belongs to the {{sink}}.
Now the problem is that is no intuitive way to distinguish between the two
(find out which operator_id is the {{map vs sink}}), just by looking at the
metrics.
Assigning names to {{map}} and {{sink}} operator does not help. Even though
these names show up in other metrics like {{numRecordsIn}}, the names does not
show up in the latency metric.
The feature request is to add {{operator_name}} along with {{operator_id}} to
the metric label so that it can be easily used in dashboards and tracking.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)