Upgrade & Secure Your Future with DevOps, SRE, DevSecOps, MLOps!
We spend hours scrolling social media and waste money on things we forget, but won’t spend 30 minutes a day earning certifications that can change our lives.
Master in DevOps, SRE, DevSecOps & MLOps by DevOps School!
Learn from Guru Rajesh Kumar and double your salary in just one year.
Ask a room of ops engineers whether they write down what an incident taught them and you will get a lot of nodding, followed by a lot of silence. The three-line lesson learned is supposed to be the easy version of that habit, yet it picks up more myths than almost any other practice. Let’s clear a few of them away, because the habit is small enough to survive being tried.
The format itself is plain. After a release, a rollback or a strange evening with the pager, you jot three lines covering the event, the lesson and the follow-up change. That is the whole thing, and it fits in a chat message.
Myth one: only major outages justify a three-line lesson learned
The loud incidents get a full review whether you like it or not. It is the quiet ones that vanish. A model that drifted a little after a weekend retrain, a correlation rule that grouped two unrelated alerts, a dashboard threshold someone nudged on a Tuesday: none of these earns a meeting, and all of them teach something.
Those small events are exactly where three lines shine. You do not need a severity level to justify writing them. If you caught yourself saying “huh, that was odd” at any point, you have enough material.
Myth two: you need a template, a tool and a meeting
Teams love to build an elaborate form before they have written a single note. Resist that. A pinned message in the team channel works. So does a plain text file in the repository next to the service it describes. The moment the process needs approval, people stop doing it.
If you want a little structure, keep the three prompts and nothing more. Happened, learned, changes. Anything beyond that is decoration, and decoration is what makes a five-minute job feel like a half-hour chore.
Myth three: three lines cannot hold anything useful
This one sounds reasonable and falls apart on contact. The limit is the feature. When you only have three lines, you have to decide which thing actually mattered, and that decision is where the learning happens. Long write-ups often bury the one sentence that would have helped you next time.
Here is a made-up example in the shape I mean. Line one: the anomaly detector paged us for a nightly batch job that had simply moved by an hour. Line two: our baseline did not know about schedule changes, only about volume. Line three: schedule changes now get tagged in the change feed before the detector retrains. Three sentences, one real fix, nothing wasted.
A useful test is whether a stranger on another shift could act on your third line without asking you a question. If they could, you wrote enough. If they would need a call, add a noun or a file name, not another paragraph. Precision beats length every time, and it keeps the note honest about what you actually know versus what you suspect.
Myth four: your tooling already remembers for you
Monitoring platforms keep a faithful history of metrics and alerts, and it is tempting to assume that history is the lesson. It is not. The graph shows that something spiked; it does not show why you changed your mind about the runbook afterward. Only a person can write the sentence that begins “next time we will”.
If you have been reading about how event noise gets tamed with smarter tooling, the piece on cutting event noise using smarter automation is a good companion. The tools surface the pattern. The three lines record what you decided to do about it.
Myth five: nobody will ever read these notes
Some of the time that is true, and it does not matter. The first reader is you in a few weeks, standing in front of a similar alert and half remembering that you have seen it before. Searching a handful of short notes takes seconds; reconstructing an evening from chat scrollback takes much longer.
The second reader is whoever joins the team next. Anyone who has tried to onboard into an AIOps practice knows how much tribal knowledge hides in people’s heads, and the walkthrough on bringing AIOps into an existing IT team makes the same point from the adoption side. Three-line notes are a cheap way to leave a trail.
Myth six: writing it down is the same as fixing it
Careful here, because this myth runs the other way. A lesson that lives only in a document is a wish. The third line, the one about what changes, is the part that needs an owner and a place in the backlog. If a note ends with “we should be more careful”, rewrite it until it names a concrete change: a new tag, a different threshold, a rollback step added to the checklist.
One habit that helps is to read the third line aloud before you post it. If you cannot picture the pull request or the config edit it points to, it is not finished.
A small aside on lessons outside your own stack
The same shape works anywhere you do a repeatable task. People who travel a lot often keep a similar scrap of text after each trip, and a site such as RankeSIM, which sets out local and travel SIM and eSIM options side by side, is the kind of reference a frequent traveler might open when drafting the “what to do differently” line for the next journey. The format does not care whether the subject is a deployment or a suitcase.
How to start this week
Pick the very next small event, however minor, and write the three lines inside a day, before the details blur. Keep them in one place your team can search. Do not edit them to sound clever. If a line feels hard to write, that is usually a sign you have found the thing worth fixing.
After a month, read the notes back in one sitting. You will spot repeats, and repeats are the cheapest source of improvement you will ever get. That payoff, more than any tool or ceremony, is why the three-line lesson learned deserves a place in an AIOps routine.