Support specifying --concurrency Envoy flag#7499
Conversation
On machines with high core count, not setting the `--concurrency` flag of Envoy results in Envoy defaulting to having as many worker threads as HW threads. In low-load scenarios, this creates very high CPU and memory usage unnecessarily. This fixes the issue by adding a concurrency flag and plumbing it through the `ContourDeployment` CRD. Signed-off-by: Moritz Sanft <58110325+msanft@users.noreply.github.com>
Signed-off-by: Moritz Sanft <58110325+msanft@users.noreply.github.com>
eae69f2 to
627def9
Compare
|
The Contour project currently lacks enough contributors to adequately respond to all PRs. This bot triages PRs according to the following rules:
You can:
Please send feedback to the #contour channel in the Kubernetes Slack |
|
Not stale |
|
The Contour project currently lacks enough contributors to adequately respond to all PRs. This bot triages PRs according to the following rules:
You can:
Please send feedback to the #contour channel in the Kubernetes Slack |
|
Not stale |
|
Any chance for a review here? cc @tsaarni @sunjayBhatia |
|
Thank you @msanft for your contribution |
|
The Contour project currently lacks enough contributors to adequately respond to all PRs. This bot triages PRs according to the following rules:
You can:
Please send feedback to the #contour channel in the Kubernetes Slack |
|
not stale |
|
There is PR coming on Envoy envoyproxy/envoy#45729 |
On machines with high core count, not setting the
--concurrencyflag of Envoy results in Envoy defaulting to having as many worker threads as HW threads. In low-load scenarios, this creates very high CPU and memory usage unnecessarily. This fixes the issue by adding a concurrency flag and plumbing it through theContourDeploymentCRD.224-thread machine without the fix:
after the fix (
concurrency: 4):