Fix profiling issue and add logs (#38)

CHANGES.md

@@ -8,8 +8,10 @@ Release Notes.
 * Support `OFF_CPU` Profiling.
 * Introduce the `BTFHub` module.
 * Update to using frequency mode to `ON_CPU` Profiling.
+* Add logs in the profiling module logical.
 
 #### Bug Fixes
+* Fix `docker` based process could not be detected.
 
 #### Issues and PR
 - All issues are [here](https://github.com/apache/skywalking/milestone/134?closed=1)

README.md

@@ -13,6 +13,8 @@ Apache SkyWalking Rover
 
 # Download
 
+Follow the [releases page](https://skywalking.apache.org/downloads/#SkyWalkingRover) or [docker image](https://hub.docker.com/r/apache/skywalking-rover) to download a release of Apache SkyWalking Rover.
+
 # Contact Us
 * Mail list: **[email protected]**. Mail to `[email protected]`, follow the reply to subscribe the mail list.
 * Join `skywalking` channel at [Apache Slack](http://s.apache.org/slack-invite). If the link is not working, find the latest one at [Apache INFRA WIKI](https://cwiki.apache.org/confluence/display/INFRA/Slack+Guest+Invites).

pkg/process/finders/kubernetes/container.go

@@ -69,6 +69,7 @@ func (c *PodContainer) CGroupID() string {
 	// delete the container runtime prefix is the cgroupid
 	cgroupID = strings.TrimPrefix(cgroupID, "containerd://")
 	cgroupID = strings.TrimPrefix(cgroupID, "dockerd://")
+	cgroupID = strings.TrimPrefix(cgroupID, "docker://")
 	return cgroupID
 }
 

pkg/profiling/task/manager.go

@@ -98,6 +98,8 @@ func (m *Manager) StartTask(c *Context) {
 	// shutdown task if exists
 	taskIdentity := c.BuildTaskIdentity()
 	if m.tasks[taskIdentity] != nil {
+		id := m.tasks[taskIdentity].TaskID()
+		log.Infof("existing profiling task: %s, so need to stop it", id)
 		if err := m.shutdownAndRemoveTask(m.tasks[taskIdentity]); err != nil {
 			log.Warnf("shutdown existing profiling task failure, so cannot to start new profiling task: %v. reason: %v", c.task.TaskID, err)
 			return
@@ -118,6 +120,7 @@ func (m *Manager) StartTask(c *Context) {
 	go func() {
 		select {
 		case <-time.After(afterRun):
+			log.Infof("the profiling task need to wait %fmin to run: %s", afterRun.Minutes(), c.TaskID())
 			m.runTask(c)
 		case <-c.ctx.Done():
 			return
@@ -126,6 +129,7 @@ func (m *Manager) StartTask(c *Context) {
 }
 
 func (m *Manager) runTask(c *Context) {
+	log.Infof("ready to starting profiling task: %s", c.TaskID())
 	var wg sync.WaitGroup
 	wg.Add(1)
 	c.runningWg = &wg
@@ -148,7 +152,7 @@ func (m *Manager) runTask(c *Context) {
 }
 
 func (m *Manager) afterProfilingStartSuccess(c *Context) {
-	log.Infof("starting the profiling task. taskId: %s, pid: %d", c.task.TaskID, c.process.Pid())
+	log.Infof("profiling task has been started. taskId: %s, pid: %d", c.task.TaskID, c.process.Pid())
 	go func() {
 		select {
 		// shutdown task when arrived task running task
@@ -223,6 +227,7 @@ func (m *Manager) flushProfilingData() error {
 		return err
 	}
 	currentMilli := time.Now().UnixMilli()
+	totalSendCount := make(map[string]int)
 	for _, t := range m.tasks {
 		data, err1 := t.runner.FlushData()
 		if err1 != nil {
@@ -234,6 +239,7 @@ func (m *Manager) flushProfilingData() error {
 			continue
 		}
 
+		totalSendCount[t.TaskID()] += len(data)
 		// only the first data have task metadata
 		data[0].Task = &profiling_v3.EBPFProfilingTaskMetadata{
 			TaskId:             t.task.TaskID,
@@ -250,6 +256,7 @@ func (m *Manager) flushProfilingData() error {
 		}
 	}
 
+	log.Infof("send profiling data summary: %v", totalSendCount)
 	_, err = stream.CloseAndRecv()
 	return err
 }

pkg/profiling/task/oncpu/runner.go

@@ -197,6 +197,8 @@ func (r *Runner) Stop() error {
 				result = multierror.Append(result, err)
 			}
 		}
+
+		close(r.stopChan)
 	})
 	return result
 }