Skip to content

Updated default timeout seconds for probes #2265

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.

Sign up for GitHub

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

Merged
kevin85421 merged 2 commits into from
Jul 28, 2024
Merged

Updated default timeout seconds for probes #2265

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fb53b1e
Updated default timeout seconds for probes
HarshAgarwal11 Jul 23, 2024
42faeda
Linting and test fix
HarshAgarwal11 Jul 28, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 18 additions & 4 deletions ray-operator/controllers/ray/common/pod.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -240,8 +240,18 @@ func DefaultWorkerPodTemplate(ctx context.Context, instance rayv1.RayCluster, wo
}

func initLivenessAndReadinessProbe(rayContainer *corev1.Container, rayNodeType rayv1.RayNodeType, creatorCRDType utils.CRDType) {
rayAgentRayletHealthCommand := fmt.Sprintf(utils.BaseWgetHealthCommand, utils.DefaultDashboardAgentListenPort, utils.RayAgentRayletHealthPath)
rayDashboardGCSHealthCommand := fmt.Sprintf(utils.BaseWgetHealthCommand, utils.DefaultDashboardPort, utils.RayDashboardGCSHealthPath)
rayAgentRayletHealthCommand := fmt.Sprintf(
utils.BaseWgetHealthCommand,
utils.DefaultReadinessProbeTimeoutSeconds,
utils.DefaultDashboardAgentListenPort,
utils.RayAgentRayletHealthPath,
)
rayDashboardGCSHealthCommand := fmt.Sprintf(
utils.BaseWgetHealthCommand,
utils.DefaultReadinessProbeFailureThreshold,
utils.DefaultDashboardPort,
utils.RayDashboardGCSHealthPath,
)

// Generally, the liveness and readiness probes perform the same checks.
// For head node => Check GCS and Raylet status.
Expand Down Expand Up @@ -279,8 +289,12 @@ func initLivenessAndReadinessProbe(rayContainer *corev1.Container, rayNodeType r
// See https://github.com/ray-project/kuberay/pull/1808 for reasons.
if creatorCRDType == utils.RayServiceCRD && rayNodeType == rayv1.WorkerNode {
rayContainer.ReadinessProbe.FailureThreshold = utils.ServeReadinessProbeFailureThreshold
rayServeProxyHealthCommand := fmt.Sprintf(utils.BaseWgetHealthCommand,
utils.FindContainerPort(rayContainer, utils.ServingPortName, utils.DefaultServingPort), utils.RayServeProxyHealthPath)
rayServeProxyHealthCommand := fmt.Sprintf(
utils.BaseWgetHealthCommand,
utils.DefaultReadinessProbeInitialDelaySeconds,
utils.FindContainerPort(rayContainer, utils.ServingPortName, utils.DefaultServingPort),
utils.RayServeProxyHealthPath,
)
commands = append(commands, rayServeProxyHealthCommand)
rayContainer.ReadinessProbe.Exec = &corev1.ExecAction{Command: []string{"bash", "-c", strings.Join(commands, " && ")}}
}
Expand Down
6 changes: 3 additions & 3 deletions ray-operator/controllers/ray/utils/constant.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -148,15 +148,15 @@ const (
LOCAL_HOST = "127.0.0.1"
// Ray FT default readiness probe values
DefaultReadinessProbeInitialDelaySeconds = 10
DefaultReadinessProbeTimeoutSeconds = 1
DefaultReadinessProbeTimeoutSeconds = 2
Copy link
Member

@andrewsykim andrewsykim Sep 4, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I ran into issues with the probe timeout in v1.1 as well.

I am thinking this probe timeout should actually be 4 or 5 seconds for the Head pod. This is because the probe for head pod runs both the agent heath check and GCS health check:

wget -T 2 -q -O- http://localhost:52365/api/local_raylet_healthz | grep success
 && wget -T 2 -q -O- http://localhost:8265/api/gcs_healthz | grep success

Which is collectively up to 4 seconds. Thoughts @kevin85421 @HarshAgarwal11

Copy link
Contributor Author

@HarshAgarwal11 HarshAgarwal11 Sep 4, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes it should be 4 to 5 seconds. Because of the OR statement, timeout might get add up. And with 2 sec I was still getting some timeouts, not as frequent as earlier. But after changing it to 5 seconds, I didn't see any timeouts, there were some failures though.

Copy link
Member

@andrewsykim andrewsykim Sep 5, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Opened #2353

Copy link
Member

@andrewsykim andrewsykim Sep 5, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Opened a separete issue to track exec probe issues: #2355

DefaultReadinessProbePeriodSeconds = 5
DefaultReadinessProbeSuccessThreshold = 1
DefaultReadinessProbeFailureThreshold = 10
ServeReadinessProbeFailureThreshold = 1

// Ray FT default liveness probe values
DefaultLivenessProbeInitialDelaySeconds = 30
DefaultLivenessProbeTimeoutSeconds = 1
DefaultLivenessProbeTimeoutSeconds = 2
DefaultLivenessProbePeriodSeconds = 5
DefaultLivenessProbeSuccessThreshold = 1
DefaultLivenessProbeFailureThreshold = 120
Expand All @@ -169,7 +169,7 @@ const (
RayAgentRayletHealthPath = "api/local_raylet_healthz"
RayDashboardGCSHealthPath = "api/gcs_healthz"
RayServeProxyHealthPath = "-/healthz"
BaseWgetHealthCommand = "wget -T 2 -q -O- http://localhost:%d/%s | grep success"
BaseWgetHealthCommand = "wget -T %d -q -O- http://localhost:%d/%s | grep success"

// Finalizers for RayJob
RayJobStopJobFinalizer = "ray.io/rayjob-finalizer"
Expand Down