Skip to content

Post-crash troubleshooting with FusionReactor On-Premise

Overview

This guide shows how to use FusionReactor On-Premise to investigate and prevent application server crashes. You’ll walk through a real-world crash scenario using tools like the Archive Viewer, Crash Protection, and JDBC Logs, and learn how to identify root causes and put safeguards in place.

Video

Key features for crash investigation

Feature Description
Archive Viewer View historical metrics, requests, and logs from before the crash
Crash Protection Automatically detect and respond to problematic requests
Request Log Analyze completed requests by duration, memory, or resource usage
JDBC Logging Capture and trace SQL statements for database-level visibility

Example crash scenario

A server crash occurs unexpectedly during peak usage. Your objective is to:

  1. Use FusionReactor On-Premise to diagnose what happened.
  2. Correlate system metrics with application activity.
  3. Prevent future downtime by configuring alerts and protective actions.

Step 1: Investigate the crash with Archive Viewer

  1. Open FusionReactor On-Premise and launch the Archive Viewer.
  2. Scroll back to the crash timeframe using log rotation timestamps.

!Screenshot

What to look for

  • In the video example, memory usage spiked at 11:49 and remained high.
  • The server crashed 14 minutes later - be sure to check the logs following the default hourly rotation.
  • By jumping back to the time of the spike, open the Request Log to examine what was happening.

Identify the trigger

  • Sort the Request Log by execution time.
  • In the video example a Store Cache request appears right as memory maxes out.
  • Other requests begin to slow down or queue, indicating resource contention.

Learn more

Archive Metrics

Step 2: Validate with logs

  • Open Server Logs for the relevant window.

  • Look for entries like:

  • java.lang.OutOfMemoryError

  • Stack traces or fatal exceptions
  • In the video scenario, an Out of Memory error is logged just before the crash, confirming the suspected cause.

!Screenshot

Step 3: Prevent recurrence with Crash Protection

Overview

Crash Protection helps you detect and manage problem requests before they crash the server.

!Screenshot

How to enable:

  1. Open the Crash Protection section in the On-Premise dashboard.
  2. Configure email settings to receive alerts.

  3. Define thresholds for:

    • Long-running or high-frequency requests
    • High CPU or memory usage

  4. Choose a response:

    • Queue or abort requests
    • Apply rules to specific endpoints

Tip

You can ignore known long-running requests to minimize false positives.

Sample alert

Here’s an example of the email notification sent when Crash Protection is triggered, including the endpoint, threshold breached, and recommended action.

!Screenshot

Learn more

Crash Protection

Optional: Enable JDBC Logging

If your crash may be database-related or you want SQL traceability:

  1. Go to JDBC → Settings in the on-prem UI.
  2. Set JDBC Logging to Enabled.

What it provides:

  • Logs every SQL statement executed by the application
  • Includes stack traces for traceability
  • Appears within the Archive Viewer for historical analysis
  • Logging levels and retention can be customized

Outcome

By using FusionReactor On-Premise, you can:

  • Rewind system state leading up to a crash.
  • Link metrics, logs, and requests to pinpoint the root cause.
  • Set up proactive rules with Crash Protection.
  • Monitor database behavior with JDBC Logging.