You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description of the problem including expected versus actual behavior:
When getting pipeline information from the node stats API, I was looking to obtain the values of pipeline.workers & pipeline.batch_size for a pipeline called "performance". However, the API only showed the general (default) values (e.g. those defined in logstash.yml).
While node_stats is likely not the place where to show configuration, it may be helpful to have.
Current Behavior
The API currently shows the information as such. The pipeline.workers and pipeline.batch_size in logstash.yml are default. However, the pipeline.workers and pipeline.batch_size for the pipeline called performance have been changed, and the API is not showing that information.
Potential Solution 1 - Add pipeline information in node_stats API
What I would expect if we are already showing "pipeline" information in "node_stats":
Potential Solution 2 - Remove 'pipelines' and 'monitoring' from node_stats API
This solution may require us to remove the pipelines and monitoring information from node_stats API. This could introduce breakage in Stack Monitoring and would mean it is no longer backwards compatible.
The text was updated successfully, but these errors were encountered:
roaksoax
changed the title
Missing 'pipeline' config information per pipeline in node stats API
Misleading pipeline information in 'node_stats' API
Dec 8, 2022
Ran into this when looking to do a calculation on flow.worker_utilization*pipeline.workers per pipeline on a monitoring dashboard. Considering that the documentation for the Node Stats API frequently mentions pipeline.workers as a multiplier for derived statistics, I think Potential Solution 1 is the way to go.
The only other way for monitoring systems to get this information requires that the monitoring developer hardcode the worker counts as constants, and update this information each time pipelines.yml is updated by an operations team. There are obvious downsides to hardcoding the number of workers in monitoring code. Ops can add new pipelines and change the number of workers, causing the monitoring to easily fall out of sync.
Logstash information:
Please include the following information:
OS version: ubuntu 20.04
Description of the problem including expected versus actual behavior:
When getting pipeline information from the node stats API, I was looking to obtain the values of pipeline.workers & pipeline.batch_size for a pipeline called "performance". However, the API only showed the general (default) values (e.g. those defined in logstash.yml).
While node_stats is likely not the place where to show configuration, it may be helpful to have.
Example
curl -XGET 'localhost:9600/_node/stats/pipelines?pretty'
Current Behavior
The API currently shows the information as such. The
pipeline.workers
andpipeline.batch_size
inlogstash.yml
are default. However, thepipeline.workers
andpipeline.batch_size
for the pipeline calledperformance
have been changed, and the API is not showing that information.Potential Solution 1 - Add pipeline information in node_stats API
What I would expect if we are already showing "pipeline" information in "node_stats":
Potential Solution 2 - Remove 'pipelines' and 'monitoring' from node_stats API
This solution may require us to remove the pipelines and monitoring information from node_stats API. This could introduce breakage in Stack Monitoring and would mean it is no longer backwards compatible.
The text was updated successfully, but these errors were encountered: