admin: add plugin lock coordination

This commit is contained in:
Chris Lu committed 2026-03-02 13:29:16 -08:00
1 parent 2644816692
commit 1e0003e283
6 files changed
+165

No files matched your search

+7
View File
@@ -0,0 +1,7 @@
package plugin
// LockManager provides a shared exclusive lock for admin-managed detection/execution.
// Acquire returns a release function that must be called when the protected work finishes.
type LockManager interface {
Acquire(reason string) (release func(), err error)
}
+10
View File
@@ -32,6 +32,7 @@ type Options struct {
SendTimeout time.Duration
SchedulerTick time.Duration
ClusterContextProvider func(context.Context) (*plugin_pb.ClusterContext, error)
LockManager LockManager
}
// JobTypeInfo contains metadata about a plugin job type.
@@ -52,6 +53,7 @@ type Plugin struct {
schedulerTick time.Duration
clusterContextProvider func(context.Context) (*plugin_pb.ClusterContext, error)
lockManager LockManager
schedulerMu sync.Mutex
nextDetectionAt map[string]time.Time
@@ -146,6 +148,7 @@ func New(options Options) (*Plugin, error) {
sendTimeout: sendTimeout,
schedulerTick: schedulerTick,
clusterContextProvider: options.ClusterContextProvider,
lockManager: options.LockManager,
sessions: make(map[string]*streamSession),
pendingSchema: make(map[string]chan *plugin_pb.ConfigSchemaResponse),
pendingDetection: make(map[string]*pendingDetectionState),
@@ -380,6 +383,13 @@ func (r *Plugin) BaseDir() string {
return r.store.BaseDir()
}
func (r *Plugin) acquireAdminLock(reason string) (func(), error) {
if r == nil || r.lockManager == nil {
return func() {}, nil
}
return r.lockManager.Acquire(reason)
}
// RunDetectionWithReport requests one detector worker and returns proposals with request metadata.
func (r *Plugin) RunDetectionWithReport(
ctx context.Context,
+15
View File
@@ -319,6 +319,21 @@ func (r *Plugin) pruneDetectorLeases(activeJobTypes map[string]struct{}) {
func (r *Plugin) runScheduledDetection(jobType string, policy schedulerPolicy) {
defer r.finishDetection(jobType)
releaseLock, lockErr := r.acquireAdminLock(fmt.Sprintf("plugin scheduled detection %s", jobType))
if lockErr != nil {
r.appendActivity(JobActivity{
JobType: jobType,
Source: "admin_scheduler",
Message: fmt.Sprintf("scheduled detection aborted: failed to acquire lock: %v", lockErr),
Stage: "failed",
OccurredAt: timeToPtr(time.Now().UTC()),
})
return
}
if releaseLock != nil {
defer releaseLock()
}
start := time.Now().UTC()
r.appendActivity(JobActivity{
JobType: jobType,