public_git/helmfile
1
0
Fork 0
mirror of https://github.com/helmfile/helmfile.git synced 2026-10-19 02:23:08 +02:00
Code Issues Projects Releases Wiki Activity

add trackLogsUntilReady flag (#2814)

Browse Source 
* add trackLogsUntilReady flag

Signed-off-by: pznamensky <kompastver@gmail.com>

* refactor(kubedog): flatten nested conditionals in trackLogsUntilReady filter

- podLogFilter.update: replace nested if-found/if-uid block with a single
  combined guard plus a no-op delete
- podLogFilter.watch: extract cutoffTrackedKind predicate and
  watchNamespacePods helper; use RWTransactionErr (kubedog's own idiom)
  instead of capturing err in plain RWTransaction closures, removing the
  'if err == nil { inf.Run() }' nesting
- TrackResources: extract newStartupLogFilter guard-clause helper so the
  printer wiring stays flat
- add table tests for the new helpers

No behavior change; race tests and golangci-lint pass.

Signed-off-by: yxxhero <aiopsclub@163.com>

---------

Signed-off-by: pznamensky <kompastver@gmail.com>
Signed-off-by: yxxhero <aiopsclub@163.com>
Co-authored-by: yxxhero <aiopsclub@163.com>
This commit is contained in:
pznamensky
2026-10-04 08:49:05 +08:00
committed by GitHub
co-authored by yxxhero
parent 000a6fa660
commit c25d16a438
22 changed files with 645 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/advanced-features.md
+27
View File
@@ -34,12 +34,39 @@ helmfile apply --track-mode kubedog --track-timeout 300 --track-logs --track-log
- **`trackMode`**: Set to `kubedog` to enable kubedog tracking, or `helm-legacy` to use Helm v4's legacy wait mode (default: `helm`)
- **`trackTimeout`**: Timeout in seconds for tracking resources (default: 300)
- **`trackLogs`**: Print logs from tracked resources during deployment
- **`trackLogsUntilReady`**: With `trackLogs: true`, stop routine pod logs once each pod first becomes ready (default: false)
- **`trackFailedLogs`**: Print collected logs only for pods that fail during deployment
To see logs only for failed pods, set `trackFailedLogs: true` on a release or use `--track-failed-logs` with `helmfile apply` or `helmfile sync`. If both log options are enabled, logs from all tracked pods are printed.
With kubedog tracking and either log option enabled, `helmfile apply` and `helmfile sync` print collected logs every 10 seconds by default. Use `--track-logs-interval` to change this interval (for example, `1s` above). The minimum is `1s`. Deployment progress is checked for changes every 10 seconds.
### Startup-Only Pod Logs
Use `trackLogsUntilReady: true` to show startup logs without continuing to print application traffic from ready pods while other replicas are still starting:
```yaml
releases:
- name: myapp
chart: ./charts/myapp
trackMode: kubedog
trackTimeout: 600
trackLogs: true
trackLogsUntilReady: true
```
Or use command-line flags:
```bash
helmfile apply --track-mode kubedog --track-timeout 600 --track-logs --track-logs-until-ready
```
Logs captured before readiness are flushed. Log output remains limited to startup if the same pod later becomes unready or fails; readiness and failure tracking continue for every pod. Job logs retain their existing behavior.
Kubernetes readiness timestamps have second precision, so logs from the readiness second are retained to avoid losing startup output.
The option defaults to `false`, so `trackLogs: true` continues printing logs throughout tracking. It only applies when `trackLogs` is enabled; `trackFailedLogs: true` alone keeps its existing failed-only behavior. When both `trackLogs` and `trackFailedLogs` are enabled, `trackLogs` takes precedence.
### Track Modes
Helmfile supports three track modes:
docs/configuration.md
+1
View File
@@ -515,6 +515,7 @@ See [Advanced Features](advanced-features.md#resource-tracking-with-kubedog) for
| `trackMode` | string | `""` | Track mode: `helm`, `helm-legacy`, or `kubedog` |
| `trackTimeout` | int | 300 | Tracking timeout in seconds |
| `trackLogs` | bool | false | Print logs from tracked resources during deployment (every 10 seconds by default; see [`--track-logs-interval`](advanced-features.md#resource-tracking-with-kubedog)) |
| `trackLogsUntilReady` | bool | false | With `trackLogs: true`, stop routine pod logs once each pod first becomes ready; readiness and failure tracking continue. Job logs are unaffected |
| `trackFailedLogs` | bool | false | Print collected logs only for pods that fail during deployment; overridden by `trackLogs` |
| `trackKinds` | list | | Whitelist of resource kinds to track |
| `skipKinds` | list | | Blacklist of resource kinds to skip |