feat(deployment): add sd_notify integration with watchdog support#25072
feat(deployment): add sd_notify integration with watchdog support#25072newtonne wants to merge 4 commits intovectordotdev:masterfrom
Conversation
github-actions
bot
added
the
domain: releasing
|
All contributors have signed the CLA ✍️ ✅ |
|
I have read the CLA Document and I hereby sign the CLA |
newtonne
changed the title
jpds
left a comment
jpds
left a comment
There was a problem hiding this comment.
Worth adding RELOADING=1 support into reload_config_from_result.
| Add systemd notify integration. Vector now sends `READY=1` when fully started, `STOPPING=1` | ||
| when beginning a graceful shutdown, and `WATCHDOG=1` pings at half the configured `WatchdogSec` | ||
| interval. The bundled `vector.service` and `hardened-vector.service` unit files are updated | ||
| to use `Type=notify`, with an optional `WatchdogSec` directive. |
There was a problem hiding this comment.
You need to add an authors: here like the other files
| let Some(duration) = sd_notify::watchdog_enabled() else { | ||
| return; | ||
| }; | ||
| let mut ticker = interval(duration / 2); |
There was a problem hiding this comment.
Also might be worth adding here:
debug!(
message = "Systemd watchdog keepalive started.",
interval_secs = ticker.as_secs_f64(),
);
|
Hey @jpds. Thanks for taking a look. I did consider this but then I discovered the
However, it doesn’t seem to be widely adopted and has its detractors so rather than opening up a can of worms, I decided to keep the scope tight with the view that this can always be added later. What do you think? |
2 participants
Summary
Adds systemd
sd_notifyintegration to Vector, enabling enhanced service lifecycle management when running under systemd withType=notify.Vector now sends:
READY=1when fully started and ready to process eventsSTOPPING=1at the beginning of graceful shutdownWATCHDOG=1keepalive pings at half the configuredWatchdogSecintervalThe bundled
vector.serviceandhardened-vector.serviceunit files are updated to useType=notify, with an optional commented outWatchdogSecdirective.See: https://www.freedesktop.org/software/systemd/man/latest/sd_notify.html
Vector configuration
No config changes required.
How did you test this PR?
Type=notify:READY=1is sent after startup is completedSTOPPING=1is sent on systemctl stopWATCHDOG=1pings are sent at the correct interval withWatchdogSecenabledType=simpleBefore
Note that systemd shows vector as started as soon as the process has forked.
After
But now systemd shows vector as started only after vector has finished starting up.
Change Type
Is this a breaking change?
Does this PR include user facing changes?
no-changeloglabel to this PR.References
Notes
@vectordotdev/vectorto reach out to us regarding this PR.pre-pushhook, please see this template.make fmtmake check-clippy(if there are failures it's possible some of them can be fixed withmake clippy-fix)make testgit merge origin masterandgit push.Cargo.lock), pleaserun
make build-licensesto regenerate the license inventory and commit the changes (if any). More details on the dd-rust-license-tool.