mirror of
https://github.com/hanzo-ml/spark-operator.git
synced 2026-10-30 20:39:07 +00:00
Venkateswarlu KomirisettiandGitHub
0741be3102
Adds a new Prometheus metric to track SparkApplication submission latency, measuring the time from application creation to the submitted state. This provides visibility into operator submission performance and helps identify bottlenecks in the job submission pipeline. The existing spark_application_start_latency_seconds metric measures time from Creation → Running state, which includes many external factors beyond the operator's control (Kubernetes scheduler delays, resource availability, image pulling, pod initialization). The new submit_latency metric measures only the operator's submission performance (Creation → Submitted state), excluding all external scheduling and resource allocation delays. Changes: - Add spark_application_submit_latency_seconds (Summary with p50, p90, p99) - Add spark_application_submit_latency_seconds_histogram (Histogram) - New CLI flag: --metrics-job-submit-latency-buckets with exponential buckets [0.5, 1, 2, 4, 8, 16, 32, 64, 128, 256] seconds - Wire through Helm chart (values, templates, tests, README) - Metric observed when SparkApplication transitions to Submitted state - Only records on first submission attempt (SubmissionAttempts == 1) - Follows same pattern as existing start latency metric Fixes kubeflow/spark-operator#2911 Signed-off-by: Venkateswarlu Komirisetti <[email protected]>
54 lines
1.9 KiB
Go
54 lines
1.9 KiB
Go
/*
|
|
Copyright 2024 The Kubeflow authors.
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
https://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
*/
|
|
|
|
package common
|
|
|
|
// Spark application metric names.
|
|
const (
|
|
MetricSparkApplicationCount = "spark_application_count"
|
|
|
|
MetricSparkApplicationSubmitCount = "spark_application_submit_count"
|
|
|
|
MetricSparkApplicationFailedSubmissionCount = "spark_application_failed_submission_count"
|
|
|
|
MetricSparkApplicationRunningCount = "spark_application_running_count"
|
|
|
|
MetricSparkApplicationSuccessCount = "spark_application_success_count"
|
|
|
|
MetricSparkApplicationFailureCount = "spark_application_failure_count"
|
|
|
|
MetricSparkApplicationSuccessExecutionTimeSeconds = "spark_application_success_execution_time_seconds"
|
|
|
|
MetricSparkApplicationFailureExecutionTimeSeconds = "spark_application_failure_execution_time_seconds"
|
|
|
|
MetricSparkApplicationSubmitLatencySeconds = "spark_application_submit_latency_seconds"
|
|
|
|
MetricSparkApplicationSubmitLatencySecondsHistogram = "spark_application_submit_latency_seconds_histogram"
|
|
|
|
MetricSparkApplicationStartLatencySeconds = "spark_application_start_latency_seconds"
|
|
|
|
MetricSparkApplicationStartLatencySecondsHistogram = "spark_application_start_latency_seconds_histogram"
|
|
)
|
|
|
|
// Spark executor metric names.
|
|
const (
|
|
MetricSparkExecutorRunningCount = "spark_executor_running_count"
|
|
|
|
MetricSparkExecutorSuccessCount = "spark_executor_success_count"
|
|
|
|
MetricSparkExecutorFailureCount = "spark_executor_failure_count"
|
|
)
|