[
https://issues.apache.org/jira/browse/NIFI-9760?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ruben Laguna updated NIFI-9760:
-------------------------------
Description:
The PrometheusReportingTask currently exposes / export the following JVM
metrics:
nifi_jvm_uptime
nifi_jvm_heap_non_usage
nifi_jvm_heap_used
nifi_jvm_gc_time
nifi_jvm_gc_runs
nifi_jvm_heap_usage
nifi_jvm_daemon_thread_count
nifi_jvm_file_descriptor_usage
nifi_jvm_thread_count
There are some notable things missing from there that are present in
nifi-api/system-diagnostics
nifi_jvm_heap_free
nifi_jvm_heap_max
nifi_jvm_heap_total
Also other non JVM stuff is missing (that is present in system diagnostics)
* contentRepositoryStorageUsage
* provenanceRepositoryStorageUsage
* flowFileRepositoryStorageUsage
I would be good to expose the same information in both places (it's a little
surprising for a newcomer that some information is in prometheus, some other in
system-diagnostics, etc)
Also it would be great to also expose all the jvm stuff under more conventional
metric names like the ones below which are taken from [AWS CloudWatch
Application Insights Java
metrics|https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/appinsights-metrics-java.html]
but are also aligned with what the
[jmx_exporter|https://github.com/prometheus/jmx_exporter] does out of the box.
- java_lang_memory_heapmemoryusage_used
- java_lang_memory_heapmemoryusage_committed
- java_lang_operatingsystem_openfiledescriptorcount
- java_lang_operatingsystem_maxfiledescriptorcount
- java_lang_operatingsystem_freephysicalmemorysize
- java_lang_operatingsystem_freeswapspacesize
- java_lang_threading_threadcount
- java_lang_threading_daemonthreadcount
- java_lang_classloading_loadedclasscount
- java_lang_garbagecollector_collectiontime_copy
- java_lang_garbagecollector_collectiontime_ps_scavenge
- java_lang_garbagecollector_collectiontime_parnew
- java_lang_garbagecollector_collectiontime_marksweepcompact
- java_lang_garbagecollector_collectiontime_ps_marksweep
- java_lang_garbagecollector_collectiontime_concurrentmarksweep
- java_lang_garbagecollector_collectiontime_g1_young_generation
- java_lang_garbagecollector_collectiontime_g1_old_generation
- java_lang_garbagecollector_collectiontime_g1_mixed_generation
- java_lang_operatingsystem_committedvirtualmemorysize
was:
The PrometheusReportingTask currently exposes / export the following JVM
metrics:
nifi_jvm_uptime
nifi_jvm_heap_non_usage
nifi_jvm_heap_used
nifi_jvm_gc_time
nifi_jvm_gc_runs
nifi_jvm_heap_usage
nifi_jvm_daemon_thread_count
nifi_jvm_file_descriptor_usage
nifi_jvm_thread_count
There are some notable things missing from there that are present in
nifi-api/system-diagnostics
nifi_jvm_heap_free
nifi_jvm_heap_max
nifi_jvm_heap_total
Also other non JVM stuff is missing (that is present in system diagnostics)
* contentRepositoryStorageUsage
* provenanceRepositoryStorageUsage
* flowFileRepositoryStorageUsage
I would be good to expose the same information in both places (it's a little
surprising for a newcomer that some information is in prometheus, some other in
system-diagnostics, etc)
Also it would be great to also expose all the jvm stuff under more conventional
metric names like
- java_lang_memory_heapmemoryusage_used
- java_lang_memory_heapmemoryusage_committed
- java_lang_operatingsystem_openfiledescriptorcount
- java_lang_operatingsystem_maxfiledescriptorcount
- java_lang_operatingsystem_freephysicalmemorysize
- java_lang_operatingsystem_freeswapspacesize
- java_lang_threading_threadcount
- java_lang_threading_daemonthreadcount
- java_lang_classloading_loadedclasscount
- java_lang_garbagecollector_collectiontime_copy
- java_lang_garbagecollector_collectiontime_ps_scavenge
- java_lang_garbagecollector_collectiontime_parnew
- java_lang_garbagecollector_collectiontime_marksweepcompact
- java_lang_garbagecollector_collectiontime_ps_marksweep
- java_lang_garbagecollector_collectiontime_concurrentmarksweep
- java_lang_garbagecollector_collectiontime_g1_young_generation
- java_lang_garbagecollector_collectiontime_g1_old_generation
- java_lang_garbagecollector_collectiontime_g1_mixed_generation
- java_lang_operatingsystem_committedvirtualmemorysize
> PrometheusReportingTask: should at least expose all the info in
> nifi-api/system-diagnostics
> --------------------------------------------------------------------------------------------
>
> Key: NIFI-9760
> URL: https://issues.apache.org/jira/browse/NIFI-9760
> Project: Apache NiFi
> Issue Type: Improvement
> Reporter: Ruben Laguna
> Priority: Major
>
> The PrometheusReportingTask currently exposes / export the following JVM
> metrics:
> nifi_jvm_uptime
> nifi_jvm_heap_non_usage
> nifi_jvm_heap_used
> nifi_jvm_gc_time
> nifi_jvm_gc_runs
> nifi_jvm_heap_usage
> nifi_jvm_daemon_thread_count
> nifi_jvm_file_descriptor_usage
> nifi_jvm_thread_count
> There are some notable things missing from there that are present in
> nifi-api/system-diagnostics
> nifi_jvm_heap_free
> nifi_jvm_heap_max
> nifi_jvm_heap_total
> Also other non JVM stuff is missing (that is present in system diagnostics)
> * contentRepositoryStorageUsage
> * provenanceRepositoryStorageUsage
> * flowFileRepositoryStorageUsage
> I would be good to expose the same information in both places (it's a little
> surprising for a newcomer that some information is in prometheus, some other
> in system-diagnostics, etc)
> Also it would be great to also expose all the jvm stuff under more
> conventional metric names like the ones below which are taken from [AWS
> CloudWatch Application Insights Java
> metrics|https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/appinsights-metrics-java.html]
> but are also aligned with what the
> [jmx_exporter|https://github.com/prometheus/jmx_exporter] does out of the
> box.
> - java_lang_memory_heapmemoryusage_used
> - java_lang_memory_heapmemoryusage_committed
> - java_lang_operatingsystem_openfiledescriptorcount
> - java_lang_operatingsystem_maxfiledescriptorcount
> - java_lang_operatingsystem_freephysicalmemorysize
> - java_lang_operatingsystem_freeswapspacesize
> - java_lang_threading_threadcount
> - java_lang_threading_daemonthreadcount
> - java_lang_classloading_loadedclasscount
> - java_lang_garbagecollector_collectiontime_copy
> - java_lang_garbagecollector_collectiontime_ps_scavenge
> - java_lang_garbagecollector_collectiontime_parnew
> - java_lang_garbagecollector_collectiontime_marksweepcompact
> - java_lang_garbagecollector_collectiontime_ps_marksweep
> - java_lang_garbagecollector_collectiontime_concurrentmarksweep
> - java_lang_garbagecollector_collectiontime_g1_young_generation
> - java_lang_garbagecollector_collectiontime_g1_old_generation
> - java_lang_garbagecollector_collectiontime_g1_mixed_generation
> - java_lang_operatingsystem_committedvirtualmemorysize
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
