Turns a ticket pile-up into one incident within minutes.
When hundreds of "no internet" tickets land at once, nobody can read them one by one. Surge catches it inside GLPI, with its own statistics.
Watch the arrivals, then drop 6 similar tickets with "Create a burst"; Surge groups them and links them to a major incident.
Main screenWhen hundreds of "no internet" tickets land at once, nobody can read them one by one. Surge catches it inside GLPI, with its own statistics.
The problem
When a switch fails, the help desk receives dozens of similar tickets within the same minutes. Each one is assigned separately, users keep reporting the same thing and the real problem is noticed late. A standard GLPI installation has no mechanism to catch this.
What I built
Surge watches the incoming stream continuously; it detects unusual pile-ups by category, location and text similarity, groups the related tickets and links them to one major incident under a parent ticket. One click adds a shared follow-up to all of them, a "we know, we're on it" notice goes up on the self-service portal, and when the parent ticket is solved all tickets close together. The on-call roster and notification rules live inside the plugin too.
Technical notes
- Detection runs on cron and only looks at new tickets on each pass
- Similarity: normalised word sets weighted with category and location
- The portal notice uses GLPI's own announcement mechanism; ticket details are never shown
- 83 end-to-end tests, run on GLPI 11 and 12 test installations
Highlights
- Statistical detection: thresholds per category and location, text-similarity percentage
- One major incident: parent ticket, problem record, shared follow-up, bulk resolution
- Portal notice and ticket banner: users see the news before opening a new ticket
- On-call and notifications: who, at what hour, through which channel
- One codebase for GLPI 11 and 12; GitHub releases with version automation