Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cost saving: Feature request to enable switch to toggle Envoy compression filter #6511

Open
chaosbox opened this issue Jun 13, 2024 · 6 comments
Labels
kind/feature Categorizes issue or PR as related to a new feature. lifecycle/needs-triage Indicates that an issue needs to be triaged by a project contributor.

Comments

@chaosbox
Copy link

Details:
Currently, Contour attaches a compression filter to envoy httpConnectionManager by default, as a result compression is enabled everywhere. While this feature is beneficial in certain scenarios, we have observed that the associated CPU cost can be prohibitively high compared to network costs. Specifically, we've found that CPU is more expensive than network bandwidth in our use case.

Rationale:
Compression is particularly useful for last-mile traffic. However, when CDNs front the ingress layer, which has dedicated bandwidth to origin servers, the network impact is negligible. In such cases, the return on investment for CPU cycles used in compression is significantly reduced. Therefore, the cost savings of disabling compression outweigh the benefits when not dealing with last-mile traffic or ingress layers meant for internal traffic.

Proposal:
We propose adding a control-plane configuration option, such as --disable-compression=true, to allow users to skip compression from the default filters. This change would provide significant cost savings in scenarios where compression is not beneficial.

Reduces CPU usage and allows for more efficient use of resources by running lean Envoy instances. We believe it will add significant value for users who, like us, need to optimize for cost and resource efficiency.

Observations from Our Cluster Testing:

  • Enabling compression increased CPU utilization by nearly three times.
    • Without compression, the same workload utilized approximately 0.6 CPU cores.
    • With compression enabled, CPU utilization spiked to 2 cores.
    • The payload per request during our test was 310KB (uncompressed) and 85KB (compressed).
      Given these results, we would prefer to run lean instances of Envoy and scale out as needed.
image
@chaosbox chaosbox added kind/feature Categorizes issue or PR as related to a new feature. lifecycle/needs-triage Indicates that an issue needs to be triaged by a project contributor. labels Jun 13, 2024
Copy link

Hey @chaosbox! Thanks for opening your first issue. We appreciate your contribution and welcome you to our community! We are glad to have you here and to have your input on Contour. You can also join us on our mailing list and in our channel in the Kubernetes Slack Workspace

@aecay
Copy link

aecay commented Jun 13, 2024

Additional context: way back in #310, the possibility was discussed to have a switch to disable the compression logic; this didn't wind up being implemented in the end. The cost savings we are discussing here (I am a colleague of @chaosbox 's) would effectively give another justification for adding the toggle (in addition to the security-related justification also mentioned on the linked ticket).

Copy link

The Contour project currently lacks enough contributors to adequately respond to all Issues.

This bot triages Issues according to the following rules:

  • After 60d of inactivity, lifecycle/stale is applied
  • After 30d of inactivity since lifecycle/stale was applied, the Issue is closed

You can:

  • Mark this Issue as fresh by commenting
  • Close this Issue
  • Offer to help out with triage

Please send feedback to the #contour channel in the Kubernetes Slack

@github-actions github-actions bot added the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Aug 13, 2024
@tsaarni tsaarni removed the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Aug 14, 2024
@geomacy
Copy link

geomacy commented Oct 9, 2024

Following initial implementation of the "disable-compression" suggestion above, it was suggested in code review to implement a choice of compression algorithms. This has now been done, see here.

@geomacy
Copy link

geomacy commented Oct 14, 2024

hi @tsaarni @davinci26 just to check, is there anything I need to do to move this issue into a "ready for review" status?

Copy link

The Contour project currently lacks enough contributors to adequately respond to all Issues.

This bot triages Issues according to the following rules:

  • After 60d of inactivity, lifecycle/stale is applied
  • After 30d of inactivity since lifecycle/stale was applied, the Issue is closed

You can:

  • Mark this Issue as fresh by commenting
  • Close this Issue
  • Offer to help out with triage

Please send feedback to the #contour channel in the Kubernetes Slack

@github-actions github-actions bot added the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Dec 17, 2024
@tsaarni tsaarni removed the lifecycle/stale Denotes an issue or PR has remained open with no activity and has become stale. label Dec 17, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
kind/feature Categorizes issue or PR as related to a new feature. lifecycle/needs-triage Indicates that an issue needs to be triaged by a project contributor.
Projects
None yet
Development

No branches or pull requests

4 participants