The ticket came back on a Monday morning, exactly as it had the Monday before. A managed-services team had closed it twice, each time with a correct fix and a polite note to the customer. Yet the same disk-space alarm kept returning, and the on-call engineer who opened it for the third time felt something between irritation and embarrassment. The problem was not technical. The problem was that nothing the team learned from the first two closures had been kept anywhere. This story walks through the diagnosis and the small habit that fixed it: a three-line lesson learned.
The symptom: the same ticket, three times
On paper the support process looked healthy. Tickets were acknowledged quickly, resolved with the right commands and closed with clear summaries. The weakness showed only when you looked across tickets. A log volume filled a server, someone cleared it, and a few days later it happened again. Each engineer solved the visible problem. Nobody recorded why it kept recurring, and the shift rotation meant the third responder never saw what the first two had done beyond a terse closing comment.
The diagnosis: fixes without memory
When the team sat down to investigate, they asked an uncomfortable question: where does a lesson live after a ticket closes? The answer was nowhere. Closing notes described the action taken, such as “cleared old logs,” but not the reasoning, the cause or the preventive step. Information existed in individual heads and evaporated with every shift change. The diagnosis, put plainly, was that the team was good at repairing and poor at remembering.
It is a common pattern in operations work. Teams measure how fast they respond and how many tickets they close, and those numbers reward speed. Speed alone, however, produces repeat visits. The same logic runs through general delivery thinking, and this site’s overview of principles that support reliable software delivery makes a similar point about feedback and learning being part of the work, not an extra.
The treatment: three lines at the end of every repeat ticket
The team did not buy a tool or launch a project. They added a three-line field to the closing form for any ticket that resembled an earlier one. Line one: what happened. Line two: what we learned about why. Line three: what we will change so it does not return. The rule was gentle by design; the field was optional for first-time issues but expected for repeats.
The first entry written under the new rule was revealing. Line one: the application server’s volume filled again because a debug logging setting remained enabled after a past investigation. Line two: nobody owned the cleanup of temporary settings, so they outlived the problem they were meant to diagnose. Line three: add rotation limits to the log configuration and add a checklist item to remove temporary debug settings when closing an investigation.
What changed after a month
Within weeks the recurring alarm stopped. More interesting was what happened to the other tickets. Engineers began searching earlier notes before starting a fix, and several found that a lesson already existed. Handovers between shifts improved because the last three lines of a ticket told the incoming engineer where things stood. The team also noticed a quieter benefit: engineers were less defensive about repeats, because the note framed a recurrence as information rather than failure.
Management saw it too. A short list of line-three actions gave them a clear view of preventive work, without any extra reporting overhead. Many of those actions pointed toward automation, which resonates with the discussion in a piece on why operations people gain from code-defined infrastructure and SRE habits, since declared configuration prevents the quiet drift that caused the original alarm.
Repeating this diagnosis with your colleagues
Start by pulling the last month of closed tickets and sorting them by similarity. Look for pairs or triples that describe the same symptom. For each cluster, ask whether the closing notes recorded a cause. If they only record actions, you have found your gap. Next, pick one cluster and have the most recent responder write a three-line lesson retroactively; this serves as a pilot and shows others the format. Finally, add the field to your closing process for repeat issues.
Keep the lines factual and keep blame out. If a line says “the engineer forgot to remove the setting,” rewrite it as “the setting had no expiry and no owner.” The second version points to a fix; the first points to a person.
Handling objections you will hear
The first objection is usually a lack of time. In reply: three lines take about two minutes, whereas a third visit to the same ticket takes far longer. Someone else will say the notes will be ignored. Prevent that by making the first step of working a repeat ticket a search of the notes, so that reading them becomes routine. A third person will say the format is too small for complex problems. Agree, and use the note as a pointer to a deeper analysis when one is warranted.
A light-touch aside on trips and planning
Staff who visit client sites can keep the note for their journeys as well. Staying online on the road is a typical worry, and RankeSIM, a site comparing SIM and eSIM choices for visitors and residents, is one place some of them begin. The lines written after the trip then say whether to repeat the approach.
The ending of the story
The Monday ticket never came back a fourth time. More importantly, the team kept the habit. Months later a new engineer joined, opened a stubborn alert and found three lines from a colleague who had since moved on. The note saved an afternoon. That is the quiet promise of the practice: a small piece of writing, done at the right moment, which turns one person’s fix into the whole team’s knowledge.