The 4 Types of Incidents as Zombies from ‘The Last of Us’
Seems like everyone has watched or is watching “The Last of Us.” This show is based on a video game of the same name. It features Pedro Pascal (from “The Mandalorian”) and his latest surrogate child, Bella Ramsey (from “Game of Thrones”). But this adventure is challenging for a plethora of reasons. Most notably, zombies. In 2003, a fungus, Cordyceps, brought on a global zombie endemic. Twenty years later, a few humans are trying to endure and survive in what’s left. Spoiler alert for anyone who hasn’t watched season one yet: it’s hard. And zombies are scary.
While incident response is rarely life or death, it can be an adrenaline spike akin to watching the show. And some of the incidents you may face have similarities to the zombies we’ve seen so far in “The Last of Us.” These incidents have a “headshot” that can help you survive against all odds.
The first zombies we see in “The Last of Us” are runners. These are fresh and may still look human compared to ones that have been infected longer. While easy to kill, there’s one factor that makes them dangerous: you never expect them. They’re novel. For those who were around for the 2003 end of the world, zombies were only a work of fiction. Nobody prepared for the end of the world (except for Bill, *sob*). For those people hanging on in 2023, runners are still jarring. They’re usually someone you know. A friend. And they change fast, as we saw at the end of episode five.
If we had to compare this one to an incident, it’d be the one that happens out of nowhere and is rare or an anomaly. The system is fine! Then it’s not and you’re thinking, “How did I miss that?” So what do you do about it? Look for the signal in the noise that tells you something is going wrong. For those infected, this could be twitching, coughing or unexpected mood swings.
There’s similar warning signs for an incident. Latency a bit high? Could be nothing. But combine it with customer support noting an increase in complaints about slowness? You may have a runner. Monitoring only gets you so far. You need to make sense of the data, both from machines and humans. Correlating that data with changes in your ecosystem can help you attack your runner before it bites you.
This is your garden variety of zombies. They’re not too difficult to kill. They don’t have any special abilities to speak of. And you can almost always expect them. Going down into the basement of an abandoned gas station? Of course you’ll find a stalker. Empty mall? Yep, you should have known, Ellie and Riley! Stalkers aren’t fun by any means, and can be deadly. More often than not, though, the average survivor can take care of a stalker. But, what happens when there’s a few stalkers all at once? Or you find 12 stalkers back to back, all in the same day? What if you’re fighting two to three stalkers every day for a year?
You can see where I’m going with this. Stalkers are like death by a thousand cuts. The more you have to tangle with them, the more dangerous they are. Like your most common incidents. They’re not fire drills, they’re annoying. And one isn’t so bad, but one every single day hurts. It takes time away from value-add work to fix something that you’ll need to fix again soon.
Automation isn’t something that Joel and Ellie can do in their world. But in our zombie-free existence, we can apply it to make incident response more efficient. For well-understood issues and incidents that happen frequently, crafting auto-remediation to resolve the problem without human intervention can immediately add time back to your day. And, it’s a great way to drive automation initiatives within the organization. Solving this small but frequent problem has a direct ROI associated with it. Leverage that to further automation initiatives for other types of incidents.
Clickers are ominous, obsessive hunters that are harder to kill. As they’re blind, they use echolocation to hunt their prey. Headshots don’t work as their heads are armored with tough fungus. They’re one of the most feared and hated types of zombies in “The Last of Us,” and it’s easy to see why. Can you imagine coming up against this thing and realizing your typical solution doesn’t work the way it should? And against a more dangerous enemy?
This one may be the hardest to correlate to an incident, because clickers seem to be almost impossible to kill in the show. Everyone’s advice? Run. Before they hear you. But with incidents, you can’t do that. So, if this zombie was an incident, it would be the one that only two or three people have seen before. You’ve heard about this issue, and it’s from deep in the tech stack. But not enough people who knew about this incident shared with the class. When it happens, it feels like a bigger issue than it is.
Like a knife to the neck of a clicker, there’s a solution to this type of incident. And success comes down to the same thing: knowledge and a plan. If you know that a clicker’s head has armor, you go for the neck. It’s close combat, but effective. And since enough people have survived clickers, the knowledge spread across the surviving population.
For an incident, the best way to fix your clickers is documentation, runbooks, and historical context. Someone knows how to resolve the problem. If they share this knowledge, teams can document the process and create a runbook for the next time this scary (but repairable) problem happens. Additionally, teams can rely on AI to surface past incident data. Look-alike incidents have lots we can learn from. This past incident data helps teams understand what worked for an incident and what didn’t. If you don’t have AI to assist, you can always scan through old retrospectives as well for this historical context. Centralizing all this information is also important so that everyone can find it. That way, you may not know how to solve every problem that happens, but you know how to find that knowledge. There’s power in that, even if there’s no perfect “headshot.”
Bloaters look more like the demogorgons in “Stranger Things” than something that was, at one point, a human. They kill most people in the vicinity either by brute force or toxic clumps of fungus that they toss in the air like grenades. We only saw one of these in “The Last of Us” so far and it made quite the impression, annihilating most of the fighting population of Kansas City. Bloaters should be avoided at ALL costs. And any signs of them should be dealt with early before the issue compounds. Remember how the zombies were filling up the tunnels and the rebels had other initiatives to take care of? Yeah, that was technical debt and someone should have fixed it.
But that’s the way it goes. You know there’s a problem, even if you don’t know exactly how it’ll manifest. Then you’ve got a major incident on your hands–a bloater. And the best and only real way to deal with these is with a coordinated, end-to-end incident response. Make sure that you understand key components of incident response such as:
- Escalation policies
- Roles and responsibilities during the incident
- Communication standards, both internal and external
- Workflows that you can trigger automatically to take the heavy lifting off responders
With these plans in place, you will be able to resolve the incident more smoothly, faster, and with less customer impact.
What zombie are you worried most about?
What’s keeping you up at night? Fear of an impending bloater, or notifications about yet another stalker? While we may not find the cure to Zombies in ‘The Last of Us,’ we can work on technology incidents and make those easier and less catastrophic for us and our customers.
PagerDuty is here to help you improve your digital operations. Whatever challenges you’re facing right now our team can help you endure and thrive, not just survive. Check out our weekly demos to learn more.