Make Actions tasks/jobs timeouts configurable by the user (#27400)

With this PR we added the possibility to configure the Actions timeouts
values for killing tasks/jobs.
Particularly this enhancement is closely related to the `act_runner`
configuration reported below:
```
# The timeout for a job to be finished.
# Please note that the Gitea instance also has a timeout (3h by default) for the job.
# So the job could be stopped by the Gitea instance if it's timeout is shorter than this.
timeout: 3h
```

---

Setting the corresponding key in the INI configuration file, it is
possible to let jobs run for more than 3 hours.

Signed-off-by: Francesco Antognazza <[email protected]>

Author: fantognazza

custom/conf/app.example.ini

@@ -2568,6 +2568,12 @@ LEVEL = Info
 ;DEFAULT_ACTIONS_URL = github
 ;; Default artifact retention time in days, default is 90 days
 ;ARTIFACT_RETENTION_DAYS = 90
+;; Timeout to stop the task which have running status, but haven't been updated for a long time
+;ZOMBIE_TASK_TIMEOUT = 10m
+;; Timeout to stop the tasks which have running status and continuous updates, but don't end for a long time
+;ENDLESS_TASK_TIMEOUT = 3h
+;; Timeout to cancel the jobs which have waiting status, but haven't been picked by a runner for a long time
+;ABANDONED_JOB_TIMEOUT = 24h
 
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

docs/content/administration/config-cheat-sheet.en-us.md

@@ -1389,6 +1389,9 @@ PROXY_HOSTS = *.github.com
 - `STORAGE_TYPE`: **local**: Storage type for actions logs, `local` for local disk or `minio` for s3 compatible object storage service, default is `local` or other name defined with `[storage.xxx]`
 - `MINIO_BASE_PATH`: **actions_log/**: Minio base path on the bucket only available when STORAGE_TYPE is `minio`
 - `ARTIFACT_RETENTION_DAYS`: **90**: Number of days to keep artifacts. Set to 0 to disable artifact retention. Default is 90 days if not set.
+- `ZOMBIE_TASK_TIMEOUT`: **10m**: Timeout to stop the task which have running status, but haven't been updated for a long time
+- `ENDLESS_TASK_TIMEOUT`: **3h**: Timeout to stop the tasks which have running status and continuous updates, but don't end for a long time
+- `ABANDONED_JOB_TIMEOUT`: **24h**: Timeout to cancel the jobs which have waiting status, but haven't been picked by a runner for a long time
 
 `DEFAULT_ACTIONS_URL` indicates where the Gitea Actions runners should find the actions with relative path.
 For example, `uses: actions/checkout@v3` means `https://github.com/actions/checkout@v3` since the value of `DEFAULT_ACTIONS_URL` is `github`.

modules/setting/actions.go

@@ -6,6 +6,7 @@ package setting
 import (
 	"fmt"
 	"strings"
+	"time"
 
 	"code.gitea.io/gitea/modules/log"
 )
@@ -18,6 +19,9 @@ var (
 		ArtifactRetentionDays int64    `ini:"ARTIFACT_RETENTION_DAYS"`
 		Enabled               bool
 		DefaultActionsURL     defaultActionsURL `ini:"DEFAULT_ACTIONS_URL"`
+		ZombieTaskTimeout     time.Duration     `ini:"ZOMBIE_TASK_TIMEOUT"`
+		EndlessTaskTimeout    time.Duration     `ini:"ENDLESS_TASK_TIMEOUT"`
+		AbandonedJobTimeout   time.Duration     `ini:"ABANDONED_JOB_TIMEOUT"`
 	}{
 		Enabled:           true,
 		DefaultActionsURL: defaultActionsURLGitHub,
@@ -82,5 +86,9 @@ func loadActionsFrom(rootCfg ConfigProvider) error {
 		Actions.ArtifactRetentionDays = 90
 	}
 
+	Actions.ZombieTaskTimeout = sec.Key("ZOMBIE_TASK_TIMEOUT").MustDuration(10 * time.Minute)
+	Actions.EndlessTaskTimeout = sec.Key("ENDLESS_TASK_TIMEOUT").MustDuration(3 * time.Hour)
+	Actions.AbandonedJobTimeout = sec.Key("ABANDONED_JOB_TIMEOUT").MustDuration(24 * time.Hour)
+
 	return err
 }

services/actions/clear_tasks.go

@@ -12,28 +12,23 @@ import (
 	"code.gitea.io/gitea/models/db"
 	"code.gitea.io/gitea/modules/actions"
 	"code.gitea.io/gitea/modules/log"
+	"code.gitea.io/gitea/modules/setting"
 	"code.gitea.io/gitea/modules/timeutil"
 )
 
-const (
-	zombieTaskTimeout   = 10 * time.Minute
-	endlessTaskTimeout  = 3 * time.Hour
-	abandonedJobTimeout = 24 * time.Hour
-)
-
 // StopZombieTasks stops the task which have running status, but haven't been updated for a long time
 func StopZombieTasks(ctx context.Context) error {
 	return stopTasks(ctx, actions_model.FindTaskOptions{
 		Status:        actions_model.StatusRunning,
-		UpdatedBefore: timeutil.TimeStamp(time.Now().Add(-zombieTaskTimeout).Unix()),
+		UpdatedBefore: timeutil.TimeStamp(time.Now().Add(-setting.Actions.ZombieTaskTimeout).Unix()),
 	})
 }
 
 // StopEndlessTasks stops the tasks which have running status and continuous updates, but don't end for a long time
 func StopEndlessTasks(ctx context.Context) error {
 	return stopTasks(ctx, actions_model.FindTaskOptions{
 		Status:        actions_model.StatusRunning,
-		StartedBefore: timeutil.TimeStamp(time.Now().Add(-endlessTaskTimeout).Unix()),
+		StartedBefore: timeutil.TimeStamp(time.Now().Add(-setting.Actions.EndlessTaskTimeout).Unix()),
 	})
 }
 
@@ -81,7 +76,7 @@ func stopTasks(ctx context.Context, opts actions_model.FindTaskOptions) error {
 func CancelAbandonedJobs(ctx context.Context) error {
 	jobs, _, err := actions_model.FindRunJobs(ctx, actions_model.FindRunJobOptions{
 		Statuses:      []actions_model.Status{actions_model.StatusWaiting, actions_model.StatusBlocked},
-		UpdatedBefore: timeutil.TimeStamp(time.Now().Add(-abandonedJobTimeout).Unix()),
+		UpdatedBefore: timeutil.TimeStamp(time.Now().Add(-setting.Actions.AbandonedJobTimeout).Unix()),
 	})
 	if err != nil {
 		log.Warn("find abandoned tasks: %v", err)