[core] [docs] (cgroups 24/n) Adding public docs for the Resource Isolation by israbbani · Pull Request #60183 · ray-project/ray

israbbani · 2026-01-15T19:21:08Z

For more information about the Resource Isolation project see #54703.

Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes.

Signed-off-by: irabbani <irabbani@anyscale.com>

gemini-code-assist

Code Review

This pull request adds comprehensive documentation for the new Resource Isolation feature using cgroup v2. The documentation is well-structured and covers requirements, usage with containers and on bare metal, API references, and troubleshooting. I've found a couple of minor inconsistencies in the code examples which I've pointed out in the review comments. Once these are addressed, this will be a great addition to the Ray documentation.

doc/source/ray-core/resource-isolation-with-cgroupv2.rst

Signed-off-by: irabbani <irabbani@anyscale.com>

doc/source/ray-core/resource-isolation-with-cgroupv2.rst

ZacAttack · 2026-01-15T22:20:01Z

doc/source/ray-core/resource-isolation-with-cgroupv2.rst

+1. System critical processes internal to Ray which are critical to node health
+2. User processes that are executing remote tasks and actors
+
+Without resource isolation, user processes can starve system processes of CPU and memory leading to node failure. Node failure can cause instability in your workload and in extreme cases lead to job failure.


Not as extreme as we'd hope lol

doc/source/ray-core/resource-isolation-with-cgroupv2.rst

Kunchd

Thanks for adding the usage docs for resource isolation! This should be very helpful for everyone.

I left a couple of little details, but the doc looks good overall.

doc/source/ray-core/resource-isolation-with-cgroupv2.rst

Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com>

Signed-off-by: irabbani <irabbani@anyscale.com>

…irabbani/ri-docs-1

Kunchd

Looks good!

Signed-off-by: irabbani <irabbani@anyscale.com>

…irabbani/ri-docs-1

edoakes

Looks great! Just a few minor comments.

We should also consider discoverability. Many times, people will not be directly searching for "resource isolation" but rather trying to solve their stability/OOM problems. So we may want to drop an example of WorkerDiedError / OOM kill errors here for SEO... and probably audit other pages to see if this should be cross-linked from anywhere. For example: https://docs.ray.io/en/latest/cluster/vms/user-guides/large-cluster-best-practices.html

doc/source/ray-core/resource-isolation-with-cgroupv2.rst

Signed-off-by: irabbani <israbbani@gmail.com>

israbbani · 2026-02-04T01:14:59Z

Looks great! Just a few minor comments.

We should also consider discoverability. Many times, people will not be directly searching for "resource isolation" but rather trying to solve their stability/OOM problems. So we may want to drop an example of WorkerDiedError / OOM kill errors here for SEO... and probably audit other pages to see if this should be cross-linked from anywhere. For example: https://docs.ray.io/en/latest/cluster/vms/user-guides/large-cluster-best-practices.html

I addressed some of this. I'll update all of our docs to x-link for SEO in a follow-up.

… ray.init(...) (#60726) Follow up from #60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com>

israbbani · 2026-02-04T18:04:40Z

@edoakes plz merge. Unrelated CI tests were failing so I had to update branch.

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: tiennguyentony <46289799+tiennguyentony@users.noreply.github.com>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com> Signed-off-by: tiennguyentony <46289799+tiennguyentony@users.noreply.github.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: tiennguyentony <46289799+tiennguyentony@users.noreply.github.com>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com> Signed-off-by: tiennguyentony <46289799+tiennguyentony@users.noreply.github.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: tiennguyentony <46289799+tiennguyentony@users.noreply.github.com>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: tiennguyentony <46289799+tiennguyentony@users.noreply.github.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com>

… ray.init(...) (#60726) Follow up from #60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: elliot-barn <elliot.barnwell@anyscale.com>

…ation (#60183) For more information about the Resource Isolation project see #54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com> Signed-off-by: elliot-barn <elliot.barnwell@anyscale.com>

… ray.init(...) (#60726) Follow up from #60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com>

…ation (#60183) For more information about the Resource Isolation project see #54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: Adel Nour <ans9868@nyu.edu>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com> Signed-off-by: Adel Nour <ans9868@nyu.edu>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: peterxcli <peterxcli@gmail.com>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com> Signed-off-by: peterxcli <peterxcli@gmail.com>

… ray.init(...) (ray-project#60726) Follow up from ray-project#60183. When not running inside privileged containers, the user will have to specify a `cgroup_path`. It makes sense for this to be a part of the public API for `ray.init(...)`. Things I'm changing 1. Promoting `cgroup_path` to a public API parameter for `ray.init` 2. Updating tests to use that parameter. 3. Running all cgroup tests on CI for all C++ and python changes. --------- Signed-off-by: irabbani <israbbani@gmail.com> Signed-off-by: peterxcli <peterxcli@gmail.com>

…ation (ray-project#60183) For more information about the Resource Isolation project see ray-project#54703. Adding public documentation for how to enable and use Resource Isolation for process isolation between system and user processes. --------- Signed-off-by: irabbani <irabbani@anyscale.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com> Signed-off-by: irabbani <israbbani@gmail.com> Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com> Signed-off-by: peterxcli <peterxcli@gmail.com>

[core] [docs] Adding public docs for the Resource Isolation.

29d56aa

Signed-off-by: irabbani <irabbani@anyscale.com>

israbbani added docs An issue or change related to documentation core Issues that should be addressed in Ray Core labels Jan 15, 2026

gemini-code-assist bot reviewed Jan 15, 2026

View reviewed changes

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Outdated Show resolved Hide resolved

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Outdated Show resolved Hide resolved

irabbani added 2 commits January 15, 2026 19:58

fixing typos

faef1f0

Signed-off-by: irabbani <irabbani@anyscale.com>

up

d2f483b

Signed-off-by: irabbani <irabbani@anyscale.com>

israbbani changed the title ~~[core] [docs] Adding public docs for the Resource Isolation.~~ [core] [docs] (cgroups 2/n) Adding public docs for the Resource Isolation Jan 15, 2026

israbbani changed the title ~~[core] [docs] (cgroups 2/n) Adding public docs for the Resource Isolation~~ [core] [docs] (cgroups 20/n) Adding public docs for the Resource Isolation Jan 15, 2026

israbbani mentioned this pull request Jan 15, 2026

[Core] Resource Isolation Milestone 1 Implementation Tracker #54703

Open

15 tasks

israbbani marked this pull request as ready for review January 15, 2026 21:56

israbbani requested a review from a team as a code owner January 15, 2026 21:56

ZacAttack reviewed Jan 15, 2026

View reviewed changes

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Outdated Show resolved Hide resolved

ZacAttack reviewed Jan 15, 2026

View reviewed changes

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Outdated Show resolved Hide resolved

ZacAttack reviewed Jan 15, 2026

View reviewed changes

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Show resolved Hide resolved

Merge branch 'master' into irabbani/ri-docs-1

41b48f3

israbbani requested a review from a team January 15, 2026 22:32

ZacAttack reviewed Jan 15, 2026

View reviewed changes

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Outdated Show resolved Hide resolved

ZacAttack reviewed Jan 15, 2026

View reviewed changes

doc/source/ray-core/resource-isolation-with-cgroupv2.rst Outdated Show resolved Hide resolved

Kunchd reviewed Jan 16, 2026

View reviewed changes

Ibrahim Rabbani and others added 4 commits January 15, 2026 17:44

Update doc/source/ray-core/resource-isolation-with-cgroupv2.rst

eb2cbb1

Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com>

Update doc/source/ray-core/resource-isolation-with-cgroupv2.rst

25d3ab5

Co-authored-by: Kunchen (David) Dai <54918178+Kunchd@users.noreply.github.com> Signed-off-by: Ibrahim Rabbani <israbbani@gmail.com>

Feedback

8663519

Signed-off-by: irabbani <irabbani@anyscale.com>

Merge branch 'irabbani/ri-docs-1' of github.com:ray-project/ray into …

670ec3d

…irabbani/ri-docs-1

Kunchd self-requested a review January 16, 2026 01:51

Kunchd approved these changes Jan 16, 2026

View reviewed changes

Ibrahim Rabbani added 3 commits January 16, 2026 12:54

Merge branch 'master' into irabbani/ri-docs-1

ab83a72

fix

411ed80

Signed-off-by: irabbani <irabbani@anyscale.com>

Merge branch 'irabbani/ri-docs-1' of github.com:ray-project/ray into …

615679e

…irabbani/ri-docs-1

edoakes reviewed Jan 17, 2026

View reviewed changes

Adding the cluster launcher docs

a4f0d3a

Signed-off-by: irabbani <israbbani@gmail.com>

edoakes enabled auto-merge (squash) February 4, 2026 01:18

Merge branch 'master' into irabbani/ri-docs-1

2d043df

github-actions bot disabled auto-merge February 4, 2026 16:50

edoakes merged commit e9ab540 into master Feb 4, 2026
6 checks passed

edoakes deleted the irabbani/ri-docs-1 branch February 4, 2026 19:03

Conversation

israbbani commented Jan 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ZacAttack Jan 15, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Kunchd left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Kunchd left a comment

Choose a reason for hiding this comment

Uh oh!

edoakes left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

israbbani commented Feb 4, 2026

Uh oh!

israbbani commented Feb 4, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

israbbani commented Jan 15, 2026 •

edited

Loading