github-mirrors/logstash
1
0
Fork 0
mirror of https://github.com/elastic/logstash.git synced 2025-04-24 14:47:19 -04:00
Code Issues Projects Releases Packages Wiki Activity

Doc: Reposition worker-utilization in doc (#16335) (#16338)

Browse source 
(cherry picked from commit eff9b540df)

Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
This commit is contained in:
github-actions[bot] 2024-07-19 13:03:07 -04:00 • committed by GitHub
parent 26d4179d88
commit 2b977ef239
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 12 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

29
docs/static/monitoring/monitoring-apis.asciidoc vendored
View file

@ -499,10 +499,6 @@ Example response:
"worker_concurrency" : {
"current": 1.973,
"lifetime": 1.721
},
"worker_utilization" : {
"current": 49.32,
"lifetime": 43.02
}
}
}
@ -540,19 +536,6 @@ Tuning a saturated pipeline to have more workers can often work to increase that
A _process_ is also considered "saturated" when its top-level `worker_concurrency` flow metric approaches the _cumulative_ `pipeline.workers` across _all_ pipelines, and similarly can be addressed by tuning the <<pipeline-stats,individual pipelines>> that are saturated.
| worker_utilization |
This is a unitless metric that indicates the percentage of available worker time being used by all plugins in a given pipeline (`duration` / (`uptime` * `pipeline.workers`).
It is useful for determining whether the pipeline has consistently-idle resources or is under resource contention.
A _pipeline_ is considered "saturated" when its `worker_utilization` flow metric approaches 100, because it indicates that all of its workers are being kept busy.
This is typically an indication of either downstream back-pressure or insufficient resources allocated to the pipeline.
Tuning a saturated pipeline to have more workers can often work to increase that pipeline's throughput and decrease back-pressure to its queue, unless the pipeline is experiencing back-pressure from its outputs.
A _pipeline_ is considered "starved" when its `worker_utilization` flow metric approaches 0, because it indicates that none of its workers are being kept busy.
This is typically an indication that the inputs are not receiving or retrieving enough volume to keep the pipeline workers busy.
Tuning a starved pipeline to have fewer workers can help it to consume less memory and CPU, freeing up resources for other pipelines.
| queue_backpressure |
This is a unitless metric representing the cumulative time spent by all inputs blocked pushing events into their pipeline's queue, relative to wall-clock time (`queue_push_duration_in_millis` / millisecond).
It is typically most useful when looking at the stats for an <<pipeline-stats,individual pipeline>>.
@ -1002,6 +985,18 @@ A positive number indicates that the queue size-on-disk is growing, and a negati
NOTE: The size of a PQ on disk includes both unacknowledged events and previously-acknowledged events from pages that contain one or more unprocessed events.
This means it grows gradually as individual events are added, but shrinks in large chunks each time a whole page of processed events is reclaimed (read more: <<garbage-collection, PQ disk garbage collection>>).
| worker_utilization |
This is a unitless metric that indicates the percentage of available worker time being used by this individual plugin (`duration` / (`uptime` * `pipeline.workers`).
It is useful for identifying which plugins in a pipeline are using the available worker resources.
A _pipeline_ is considered "saturated" when `worker_utilization` approaches 100, because it indicates that all of its workers are being kept busy.
This is typically an indication of either downstream back-pressure or insufficient resources allocated to the pipeline.
Tuning a saturated pipeline to have more workers can often work to increase that pipeline's throughput and decrease back-pressure to its queue, unless the pipeline is experiencing back-pressure from its outputs.
A _pipeline_ is considered "starved" when `worker_utilization` approaches 0, because it indicates that none of its workers are being kept busy.
This is typically an indication that the inputs are not receiving or retrieving enough volume to keep the pipeline workers busy.
Tuning a starved pipeline to have fewer workers can help it to consume less memory and CPU, freeing up resources for other pipelines.
|===
[discrete]