NotificationsYou must be signed in to change notification settings
Fork6.1k
Star75.1k

nd_watcher#20103

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.

By clicking “Sign up for GitHub”, you agree to ourterms of service andprivacy statement. We’ll occasionally send you account related emails.

Already on GitHub?Sign in to your account

Jump to bottom

Open

ktsaou wants to merge4 commits intonetdata:master

base:master

Choose a base branch

fromktsaou:feature/parent-process-monitoring

Open

nd_watcher#20103

ktsaou wants to merge4 commits intonetdata:masterfromktsaou:feature/parent-process-monitoring

Conversation

Copy link

Member

ktsaou commentedApr 9, 2025•
edited
Loading

Add parent process monitoring to update status file on hard kills.

This commit adds a parent-child process architecture to improve status file accuracy when Netdata is terminated abruptly. The implementation:

Creates a parent process at startup that monitors the Netdata child process
Parent process changes its name to "nd_watcher" to avoid being killed by pkill
When the child process is killed abruptly, the parent updates the status file
Adds EXIT_REASON_HARD_KILLED flag for processes killed without a recognizable signal
Forwards termination signals from parent to child
Centralizes status file loading with daemon_status_file_load_latest()

With this, the process tree looks like this:

 535187 pts/4    Ss+    0:00  |       \_ nd_watcher 535204 pts/4    Sl+   12:20  |           \_ /usr/bin/netdata -D 535206 pts/4    S+     0:00  |               \_ spawn-plugins       535820 pts/4    S+     0:00  |                   \_ bash /usr/libexec/netdata/plugins.d/tc-qos-helper.sh 1 535822 pts/4    Sl+    0:05  |                   \_ /usr/libexec/netdata/plugins.d/go.d.plugin 1 536149 pts/4    S+     0:46  |                   |   \_ /bin/intel_gpu_top -J -s 900 535862 pts/4    Sl+    0:01  |                   \_ /usr/libexec/netdata/plugins.d/systemd-journal.plugin 1 535871 pts/4    S+     0:10  |                   \_ /usr/libexec/netdata/plugins.d/slabinfo.plugin 1 535894 pts/4    Sl+    0:00  |                   \_ /usr/libexec/netdata/plugins.d/network-viewer.plugin 1 535968 pts/4    S+     0:00  |                   |   \_ spawn-setns                                            535904 pts/4    S+     0:00  |                   \_ /usr/libexec/netdata/plugins.d/perf.plugin 1 all 535911 pts/4    Sl+    0:48  |                   \_ /usr/libexec/netdata/plugins.d/apps.plugin 1 535922 pts/4    S+     0:01  |                   \_ /usr/libexec/netdata/plugins.d/cups.plugin 1 535935 pts/4    Sl+    0:05  |                   \_ /usr/libexec/netdata/plugins.d/debugfs.plugin 1 535954 pts/4    S+     0:00  |                   \_ /usr/libexec/netdata/plugins.d/nfacct.plugin 1

Add parent process monitoring to update status file on hard kills

44b36fe

This commit adds a parent-child process architecture to improve status file accuracy when Netdata is terminated abruptly. The implementation:1. Creates a parent process at startup that monitors the Netdata child process2. Parent process changes its name to "nd_watcher" to avoid being killed by pkill3. When the child process is killed abruptly, the parent updates the status file4. Adds EXIT_REASON_HARD_KILLED flag for processes killed without a recognizable signal5. Forwards termination signals from parent to child6. Centralizes status file loading with daemon_status_file_load_latest()