Every page in the Web Emergency Playbook ends the same way. After the domain is renewed, the site restored, the intruder evicted, or the account recovered, the final step on the page is the same five words: write the incident log entry. The step is easy to skip. The fire is out, the adrenaline is fading, and the last thing anyone wants to do at seven in the evening is write a paragraph about it. Yet that paragraph is the cheapest insurance an organization can buy, and this post is the case for keeping it.

The purpose of an incident log is to convert the worst web day into documentation: a dated record of what happened, how it got in, what was done, and what changed. Understanding why it matters is the difference between a log that gets written and one that gets promised. By the end you should know what an entry contains, where it lives, who reads it, and why the organizations that get through incidents well are nearly always the ones that wrote things down as they went.

What an incident log is

An incident log is a running, dated record of every security or continuity event the organization's web presence has had: outages, compromises, lockouts, expired domains, broken updates, phished passwords. One entry per event, written within a day of closing it, kept in one place. It is not a ticketing system, and it needs no software. A tab in the decisions log, a page in the operations folder, or a shared document titled plainly enough that a successor would open it serves fine.

Chapter 8 of the book lays out the incident response sequence, and its final step is the log entry. The reason it sits last is not that it matters least. The sequence is written for a shaken person, and the shaken person needs to contain, rotate, call the host, and restore before anything else. The log is where the shaking stops and the professional work begins.

What one entry contains

The playbooks each ask for slightly different fields, because each fire leaves different evidence, and read together they describe the standard entry. From the hacked-site page: what happened, when, how it got in, what was done, and what changed. From the site-down page: what failed, which layer, how long the site was down, the cause, and the fix. From the locked-out page: how long recovery took and what evidence the provider's support asked for. From the mailbox page Tuesday covered: the entry vector, the dwell time between first entry and eviction, and the actions taken. The triage page adds the rule for a day with several fires at once: keep one log all day, with times, actions, and findings in one place, then write one entry for the cluster with the real root cause and the total timeline.

So a complete entry has seven parts. The date and time of the first sign, and who noticed. What was observed, with the screenshots the playbooks told you to take. How the problem got in, or "unknown" if the host's logs could not say. The timeline of actions, with times. The time to resolution. The root cause. And what changed afterward: the plugin removed, the second administrator named, the rule added to the calendar. Most entries run under two hundred words. The discipline is in writing them at all.

Who reads it

Four readers make the log worth its cost, and only one of them is you.

The first is the next incident. Reinfection within days of a restore is the signature of a missed backdoor, and the only way to recognize a repeat is a record of the original. A host that broke the site twice in one quarter is a pattern, and the playbooks are plain about the difference: one bad afternoon is weather, a pattern is a landlord problem. Without the log, every incident is the first one.

The second is leadership. Chapter 7 recommends a quarterly three-sentence note drawn from the maintenance calendar: what ran, what it caught, what it cost in hours. The incident log is where the middle sentence comes from. "One broken payment link caught and fixed before anyone reported it" is a line that earns budget, and it exists only if someone wrote it down the day it happened. Maintenance is invisible by design; nobody thanks anyone for the outage that did not happen. The log is how the non-event becomes known work.

The third is your successor. The hit-by-a-bus file from Chapter 2 points to the inventory and the vault. The incident log is the part of the operations folder that tells the next person what actually goes wrong here, how often, and what fixed it. A successor who reads three years of entries in an afternoon knows the site better than most of the people who built it.

The fourth is the one nobody wants to meet: the attorney, the insurer, or the regulator. If an incident ever involves personal data, customer records, donor information, form submissions, disclosure duties may apply, and the first question any professional will ask is what happened and when. A dated, factual log written at the time is evidence of good faith. A reconstruction from memory three weeks later is a story. Chapter 17 of the book covers the legal side; the log is what you bring to that conversation.

Where the log lives

The same rule that governs the hit-by-a-bus file governs this one: it must be findable by someone who is not you, and it must not contain passwords. Keep it in organizational storage rather than on a personal laptop, and name its location in the same place leadership already looks, such as the board minutes or the operations folder's index. Two named people beyond you should know where it is.

Two conventions keep it honest. Write entries in plain, factual language with no speculation; "we may have been hacked by" sentences have a way of outliving their accuracy, and the log should contain only what can be supported. And date everything: the first sign, each action, the resolution, and the entry itself. Timestamps are the whole value of the thing.

The near-miss belongs in it too

The log is not only for disasters. The update that broke the layout for an hour on a Tuesday, the renewal notice that arrived for a card that had expired, the phishing email a colleague reported within ten minutes: all of these are entries. Near-misses are the cheapest lessons the organization will ever get, and a log that records them shows, over a year, exactly where the defenses are being tested. The colleague who reported the click fast deserves to see that number written down, because that speed is the organization's real security posture.

The mission tie-back

Security Month has spent six weeks adding locks, and all of them matter. The incident log is the one piece of the system that makes every other piece better over time, because it is the only part that learns. An organization that can say what happened, when, and what changed has done something most organizations its size never manage: it has turned a bad day into an asset. That is what cheap insurance looks like at this scale, and the premium is a paragraph.

Get the triage page while nothing is wrong. The Web Emergency Playbook's "Everything Is On Fire" page ends with the one-log rule for a multi-fire day, and every playbook in the kit closes with the entry described above. Print it, fill in the contact lines, and put it where the next incident's first timestamp will be written.

Categories

Governance  Management