Failure Detection
Configure real-time build failure alerts and root-cause analysis with BuildButler.
BuildButler monitors every build in real time and can alert you within seconds of a failure — before anyone notices the red ball in Jenkins.
How it works
- The collector detects a completed build with
result != SUCCESS. - It fetches the console log and runs it through the Log Analyzer.
- The analyser extracts error signatures (stack traces, exit codes, known patterns).
- A notification is dispatched with the failure summary.
Enabling failure alerts
Go to Settings → Alerts → New Rule and configure:
Root-cause categories
The log analyser classifies failures into categories:
| Category | Example patterns |
|---|---|
| Compilation | error TS2304, cannot find symbol, SyntaxError |
| Test failure | FAIL src/, AssertionError, expected X but got Y |
| Dependency | npm ERR!, Could not resolve dependencies, 401 Unauthorized |
| Infrastructure | OOMKilled, No space left on device, Connection refused |
| Timeout | Build timed out, deadline exceeded |
| Flaky | Test passed on retry — BuildButler tracks flaky rate per test |
Flaky test tracking
When a job is configured with retry-on-failure, BuildButler compares the test results between the failed and passed runs. Tests that flip between pass and fail are flagged as flaky with a confidence score.
Alert channels
BuildButler supports the following notification channels:
- Slack — via incoming webhook or the BuildButler Slack app
- Email — SMTP or SendGrid
- Microsoft Teams — via incoming webhook
- PagerDuty — incident creation via Events API v2
- Webhook — generic HTTP POST with JSON payload
See Notifications for detailed setup instructions for each channel.