Skip to content

Liveness agent state #9673

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

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
+65 −82
Open

Liveness agent state #9673

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
b4d0816
fix(tests): update liveness/readiness test cases to assert status cod…
nkvoll Sep 1, 2025
6a26207
fix: correct order of fields in LivenessFailConfig for degraded state
nkvoll Sep 1, 2025
c5dbf9d
fix: remove unnecessary check for coordinator mode in liveness handle…
nkvoll Sep 1, 2025
beff418
fix: add unhealthy coordinator state handling in liveness handler
nkvoll Sep 1, 2025
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
13 changes: 6 additions & 7 deletions internal/pkg/agent/application/monitoring/liveness.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@ import (
"github.com/elastic/elastic-agent-client/v7/pkg/client"

"github.com/elastic/elastic-agent/internal/pkg/otel/otelhelpers"
agentclient "github.com/elastic/elastic-agent/pkg/control/v2/client"
)

const formValueKey = "failon"
Expand Down Expand Up @@ -44,7 +45,7 @@ func handleFormValues(req *http.Request) (LivenessFailConfig, error) {
case "failed":
return LivenessFailConfig{Degraded: false, Failed: true, Heartbeat: true}, nil
case "degraded":
return LivenessFailConfig{Failed: true, Degraded: true, Heartbeat: true}, nil
return LivenessFailConfig{Degraded: true, Failed: true, Heartbeat: true}, nil
case "heartbeat", "":
return defaultUserCfg, nil
default:
Expand Down Expand Up @@ -76,12 +77,10 @@ func livenessHandler(coord CoordinatorState) func(http.ResponseWriter, *http.Req
return fmt.Errorf("error handling form values: %w", err)
}

// if user has requested `coordinator` mode, just revert to that, skip everything else
if !failConfig.Degraded && !failConfig.Failed && failConfig.Heartbeat {
Copy link
Member Author

Choose a reason for hiding this comment

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

I removed this part as it's already covered by line 70-73, without the encapsulating if-statement.

if !isUp {
w.WriteHeader(http.StatusServiceUnavailable)
return nil
}
unhealthyState := (failConfig.Failed && state.State == agentclient.Failed) || (failConfig.Degraded && state.State == agentclient.Degraded)
if unhealthyState {
w.WriteHeader(http.StatusInternalServerError)
return nil
}

unhealthyComponent := false
Expand Down
113 changes: 46 additions & 67 deletions internal/pkg/agent/application/monitoring/liveness_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,14 +13,14 @@ import (
"time"

"github.com/open-telemetry/opentelemetry-collector-contrib/pkg/status"
"go.opentelemetry.io/collector/component/componentstatus"

"github.com/stretchr/testify/require"
"go.opentelemetry.io/collector/component/componentstatus"

"github.com/elastic/elastic-agent-client/v7/pkg/client"
"github.com/elastic/elastic-agent/internal/pkg/agent/application/coordinator"
"github.com/elastic/elastic-agent/pkg/component"
"github.com/elastic/elastic-agent/pkg/component/runtime"
agentclient "github.com/elastic/elastic-agent/pkg/control/v2/client"
)

type mockCoordinator struct {
Expand All @@ -41,15 +41,19 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
defer cancel()
testCases := []struct {
name string
coord mockCoordinator
coord CoordinatorState
expectedCode int
liveness bool
failon string
}{
{
name: "healthy-nocoord",
coord: nil,
expectedCode: 200,
failon: "heartbeat",
},
{
name: "healthy",
expectedCode: 200,
liveness: true,
failon: "heartbeat",
},
{
Expand All @@ -72,7 +76,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 200,
liveness: true,
failon: "heartbeat",
},
{
Expand All @@ -95,7 +98,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 200,
liveness: true,
failon: "heartbeat",
},
{
Expand All @@ -118,7 +120,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 500,
liveness: true,
failon: "degraded",
},
{
Expand All @@ -141,34 +142,10 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 200,
liveness: true,
failon: "failed",
},
{
name: "degraded-liveness-off",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
Components: []runtime.ComponentComponentState{
{
LegacyPID: "2",
State: runtime.ComponentState{State: client.UnitStateDegraded},
Component: component.Component{
ID: "test-component",
InputSpec: &component.InputRuntimeSpec{
BinaryName: "testbeat",
},
},
},
},
},
},
expectedCode: 200,
liveness: false,
failon: "degraded",
},
{
name: "healthy",
name: "healthy-components",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
Expand All @@ -187,7 +164,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 200,
liveness: true,
failon: "degraded",
},
{
Expand All @@ -210,7 +186,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 503,
liveness: true,
failon: "heartbeat",
},
{
Expand All @@ -233,7 +208,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 503,
liveness: true,
failon: "heartbeat",
},
{
Expand All @@ -256,7 +230,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 503,
liveness: true,
failon: "degraded",
},
{
Expand All @@ -279,7 +252,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 503,
liveness: true,
failon: "degraded",
},
{
Expand All @@ -302,30 +274,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 503,
liveness: true,
failon: "degraded",
},
{
name: "healthy-liveness-off",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
Components: []runtime.ComponentComponentState{
{
LegacyPID: "5",
State: runtime.ComponentState{State: client.UnitStateHealthy},
Component: component.Component{
ID: "test-component3",
InputSpec: &component.InputRuntimeSpec{
BinaryName: "testbeat",
},
},
},
},
},
},
expectedCode: 200,
liveness: false,
failon: "degraded",
},
{
Expand Down Expand Up @@ -358,11 +306,10 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 500,
liveness: true,
failon: "degraded",
},
{
name: "healthy-liveness-off-otel",
name: "healthy-otel",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
Expand All @@ -380,7 +327,6 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 200,
liveness: false,
failon: "degraded",
},
{
Expand All @@ -402,7 +348,39 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
},
},
expectedCode: 500,
liveness: true,
failon: "degraded",
},
{
name: "unhealthy-failed-coordinator",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
State: agentclient.Failed,
},
},
expectedCode: 500,
failon: "failed",
},
{
name: "unhealthy-degraded-coordinator",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
State: agentclient.Degraded,
},
},
expectedCode: 500,
failon: "degraded",
},
{
name: "healthy-degraded-coordinator",
coord: mockCoordinator{
isUp: true,
state: coordinator.State{
State: agentclient.Failed,
},
},
expectedCode: 500,
failon: "degraded",
},
}
Expand All @@ -418,8 +396,9 @@ func TestLivenessProcessHTTPHandler(t *testing.T) {
require.NoError(t, err)
res, err := http.DefaultClient.Do(req)
require.NoError(t, err)
res.Body.Close()
defer res.Body.Close()

require.Equal(t, test.expectedCode, res.StatusCode)
})
}

Expand Down
21 changes: 13 additions & 8 deletions internal/pkg/agent/application/monitoring/readiness_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,24 +19,27 @@ func TestReadinessProcessHTTPHandler(t *testing.T) {
defer cancel()
testCases := []struct {
name string
coord mockCoordinator
coord CoordinatorState
expectedCode int
liveness bool
}{
{
name: "healthy-nocoord",
coord: nil,
expectedCode: 200,
liveness: true,
},
{
name: "healthy",
name: "healthy",
coord: mockCoordinator{
isUp: true,
},
expectedCode: 200,
liveness: true,
},
{
name: "unhealthy",
name: "unhealthy",
coord: mockCoordinator{
isUp: false,
},
expectedCode: 503,
liveness: false,
},
}

Expand All @@ -50,7 +53,9 @@ func TestReadinessProcessHTTPHandler(t *testing.T) {
require.NoError(t, err)
res, err := http.DefaultClient.Do(req)
require.NoError(t, err)
res.Body.Close()
defer res.Body.Close()

require.Equal(t, test.expectedCode, res.StatusCode)
})
}

Expand Down
Loading