fix: preserve server schedule failures through tee pipeline

[Gabrielgvl]

Summary

Server schedules currently wrap the remote script with tee -a <log> but only use set -e.

That means the pipeline exit status comes from tee, not from the scheduled script. If the remote script exits non-zero but tee succeeds, Dokploy can still mark the deployment as done.

This change enables pipefail for the scheduleType === "server" execution path so non-zero script exits propagate correctly.

Reproduction

With the current code, a server schedule like this can still finish as done:

#!/usr/bin/env bash
set -euo pipefail
exit 42

The current wrapper is:

set -e
bash -c /path/to/script.sh 2>&1 | tee -a /path/to/log || {
  echo "❌ Command failed" >> /path/to/log
  exit 1
}
echo "✅ Command executed successfully" >> /path/to/log

Because the pipeline does not use pipefail, tee can return 0 and hide the script failure.

Fix

Change the wrapper to:

set -euo pipefail

so the pipeline fails when the scheduled script fails.

Validation

• pnpm exec biome check packages/server/src/utils/schedules/utils.ts

Greptile Summary

This PR fixes a real bug in the server schedule type where shell pipeline exit codes were silently dropped. The wrapper command used set -e but not pipefail, meaning the pipeline bash -c script.sh 2>&1 | tee -a log would return tee's exit code (typically 0) rather than the script's exit code. If the scheduled script exited non-zero, the || { echo "❌ Command failed"; exit 1; } guard would never fire and the deployment would be incorrectly marked as done.

The one-line fix — changing set -e to set -euo pipefail — correctly addresses this. With pipefail, a non-zero exit from any command in a pipeline propagates as the pipeline's exit status, causing the || handler to trigger and the deployment to be correctly marked as error.

Key observations:

• set -euo pipefail is a standard bash idiom; -u (nounset) is a safe addition since all variables (${deployment.logPath}, ${fullPath}) are interpolated by TypeScript before the shell sees them.
• The other schedule paths (application/compose remote and dokploy-server) don't use tee in a pipeline, so they don't have the same issue and don't require the same fix.
• The change is minimal, targeted, and fixes the exact failure mode described in the PR.

Confidence Score: 5/5

• Safe to merge — the fix is minimal, targeted, and correctly resolves a real deployment status bug in the server schedule path.
• Single-line change from set -e to set -euo pipefail, which is the canonical fix for the described tee-pipeline exit-code problem. No regressions are introduced: all variable substitutions are TypeScript-side, -u is safe here, and the || guard still works correctly with pipefail active. No other files are affected.
• No files require special attention.

_{Reviews (1): Last reviewed commit: "fix: preserve server schedule failures t..." | Re-trigger Greptile}

_{(2/5) Greptile learns from your feedback when you react with thumbs up/down!}