Two hundred tickets are usually one bug

How I read a ticket queue for the shape underneath it instead of working through it in order.

Published
Reading time
5 min
Author
Ahmet Mustafa Yılmaz

A queue that suddenly has two hundred items in it is not two hundred problems. It is somewhere between one and four problems described two hundred different ways by people who do not share vocabulary.

Working the queue in order finds this out on ticket sixty. Reading the queue first finds it in ten minutes.

What I read for

Time of first arrival. Sort by creation and look at where the rate changes. A queue that was normal until 08:40 and then was not tells you when the cause started, and something happened at 08:40 that you can go and look at.

The population, not the description. Users describe symptoms in their own language and the same fault produces five different descriptions. What they have in common is more informative: same office, same laptop model, same Office build, same group. One of those columns is almost always the answer.

Who is not in the queue. The absence is as informative as the presence. If nobody from the third floor has reported it, the third floor is different in some way, and that difference is a strong hint at the cause.

Why the queue resists this

Ticket systems are built for throughput. They present a list, they measure how fast you clear it, and every incentive points at working items rather than looking across them. The reporting is per ticket, so the pattern that would let you close a hundred at once is not a thing the tool shows you.

The practical answer is to export and look at it somewhere the tool cannot impose its shape. I have found more root causes in a spreadsheet of ticket metadata than in the tickets themselves.

The one that stays with me

Two hundred users reporting a slow shared drive over three days. Every ticket individually was plausible and got a plausible individual answer. The pattern was that all of them had logged in after a specific change window, and the users who had not logged out since before it were fine.

One cause, three days of work spent on the symptoms. Now I read the queue before I work it, every time, and the ten minutes has never once been wasted.

Read next

Tell me what is breaking.

Migration work, an estate that has outgrown its last design, or a question about the toolkit. All of it is welcome.

Follow the writing

Prefer RSS? Subscribe to the feed.