[DONOT_MERGE] node-density-cni on 500 nodes #641
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Below changes are required to run node-density-cni on 500+ nodes 1) requestTimeout is set to 60 sec to avoid Client.Timeout errors while checking for created objects 2) Better to use metrics-aggregated.yml instead of metrics.yml which will have reduced/aggregated metrics. kube-burner looks for metrics.yml file in current directory (or generates if not found) and uses that for metrics. This patch adds metrics.yml in the e2e workload directory which will have only required metrics. 3) At large scale (500 node+), prometheous is failing to scrape containerCPU-AggregatedWorkers, containerMemory-AggregatedWorkers, nodeCPU-AggregatedWorkers, nodeMemoryUtilization-AggregatedWorkers, podStatusCount and podDistribution. So removed them from the metrics.yml 4) we manually need to reduce node count after test is finished and run 'kube-burner index' to scrape containerCPU-AggregatedWorkers and containerMemory-AggregatedWorkers metrics if needed.
Label workers with ovnic and then scrape metrics from only these workers. node-desnity-cni on 500 nodes runs for 2 hours 15 minutes. Scraping metrics from 500 nodes for the duration of 2 hours 15 minutes is overkill. So we scrape from only 10 worker nodes if the worker node count is more than 120.
Scale the machineset to the desired count before running kube-burner. run.sh accepts the number of workers to scale for each avaialability zone in region as env variable. Currently this patch is hard coded to support us-west-2 region with 4 avaialaility zones. For example, if user wants to scale US_WEST_2A to 143 nodes and US_WEST_2B to 188 nodes, pass them to run.sh like below US_WEST_2A=143 US_WEST_2B=188 WORKLOAD=node-desnity-cni ./run.sh note: it scales nodes 50 at a time
venkataanil
force-pushed
the
large_scale
branch
from
4f08d18
to
9aa5ebf
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Below changes are required to run node-density-cni on 500+ nodes 1) requestTimeout is set to 60 sec to avoid Client.Timeout errors while checking for created objects
2) Better to use metrics-aggregated.yml instead of metrics.yml which will have reduced/aggregated metrics. kube-burner looks for metrics.yml file in current directory (or generates if not found) and uses that for metrics. This patch adds metrics.yml in the e2e workload directory which will have only required metrics. 3) At large scale (500 node+), prometheous is failing to scrape containerCPU-AggregatedWorkers, containerMemory-AggregatedWorkers, nodeCPU-AggregatedWorkers, nodeMemoryUtilization-AggregatedWorkers, podStatusCount and podDistribution. So removed them from the metrics.yml 4) we manually need to reduce node count after test is finished and run 'kube-burner index' to scrape containerCPU-AggregatedWorkers and containerMemory-AggregatedWorkers metrics if needed.
Type of change
Description
Related Tickets & Documents
Checklist before requesting a review
Testing