How much did that outage cost?

People like to put dollar values on outages. There’s no true answer to the question of how much an outage costs an organization. If your company is transaction-based, you can estimate how many transactions were missed, but there are all sorts of other factors that you could decide to model. (Are those transactions really lost, or will people come back later? Does it impact the public’s perception of the organization? What if your business isn’t transaction-based?). If you ask John Allspaw, he’ll tell you that incidents can provide benefits to an organization in addition to costs.

Putting all of that aside for now, one question around incident cost that I think is interesting is the perceived cost within the organization. How costly does leadership feel that this incident was?

Here’s a proposed approach to try and capture this. After an incident, go to different people in leadership, and ask them the following question:

Imagine I could wave a magic wand, and it would alter history to undo the incident: it will be as if the incident never happened. However, I’ll only do this if you pay me money out of your org’s budget. How much are you willing to pay me to wave the wand?

I think this would be an interesting way to convey to the organization how leadership perceived the incident.

“What could we have done differently?”

During incident retrospective meetings, I’ve often heard someone ask: “What could we have done differently?” I don’t like this question, and so I never ask it.

A world that never was

I am a firm believer in the idea that the best way to get better at dealing with incidents is to understand how incidents actually happen. After an incident happens, I focus all of my energies on the understanding aspect, because the window of opportunity for studying the incident closes quickly.

Asking “what could we have done differently?” can’t teach us anything about how the incident happened, because it’s asking us to imagine an alternate reality where events unfolded differently. You can’t get a better understanding of why an incident responder took action X by imagining a world where the responder took action Y.

Instead of asking how it could have unfolded differently, you’ll learn a lot more about the incident if you try to understand the frame of mind of the incident responders. What did they see? What did they know at the time? What was confusing to them?

The future, not the past

I believe the question is well-intended, to help us prevent the incident from recurring. In that case, I think a better question would be something along the lines of: “If we encounter similar symptoms in a future incident, what actions should we take?” This sounds like the same question, but it’s not:

“If we encounter similar symptoms” introduces uncertainty into the exercise – the future incident may look like the last one, but it might be different with the same symptoms! When we ask about doing things differently in the past, it’s all too easy to forget about this uncertainty.

Uncertainty is one of the defining characteristics of an incident. The system is behaving in an unexpected way, and we don’t understand why! When we look back on an incident, we should focus on this uncertainty rather than elide it.

Another reason that imagining future scenarios is better that counterfactuals about past scenarios is that our system in the future is different from the one in the past. For example:

  • You may have made changes to the system in the wake of the last incident that prevents the incident from recurring in exactly the same way as before, so the question turns out to be moot.
  • You may have improved the operability of your system in some way (e.g., added an admin interface so you can make an API call instead of poking at the database), so that you have new actions you can take in the future that you couldn’t take in the past.

While I still probably wouldn’t ask this question (I want to spend all of my energy understanding the incident), I think it’s a much better question, because it gives us practice at anticipating future incidents.

OOPS writeups

A couple of people have asked me to share how I structure my OOPS write-ups. Here’s what they look like when I write them. This structure in this post is based on the OOPS template that has evolved over time inside of Netflix, with contributions from current and former members of the CORE team.

My personal outline looks like this (the bold sections are the ones that I include in every writeup)

  • Title
  • Executive summary
  • Background
  • Narrative description
    • Prologue
    • The trigger
    • Impact
    • Epilogue
  • Contributors/enablers
  • Mitigators
  • Risks
  • Challenges in handling

Title: OOPS-NNN: How we got here

Every OOPS I write up has the same title, “how we got here”. However, the name of the Google doc itself (different from the title) is a one-line summary, for example: “Server groups stuck in ‘deploying’ state”.

Executive summary

I start each write-up with a summary section that’s around three paragraphs. I usually try to capture:

  • When it happened
  • The impact
  • Explanation of the failure mode
  • Aspects about this incident that were particularly difficult

On <date>, from <start time> to <end time>, users were unable to <symptom>

The failure mode was triggered by an unintended change in <service> that led to <surprising behavior>.

The issue was made more difficult to diagnose/remediate due to a number of factors:

  • <first factor>
  • <second factor>

I’ll sometimes put the trigger in the summary, as in the example above. It’s important not to think of the trigger as the “root cause”. For example, if an incident involves TLS certificates expiring, then the trigger is the passage of time. I talk more about the trigger in the “narrative description” section below.


It’s almost always the case that the reader will need to have some technical knowledge about the system in order to make sense of the incident. I often put in a background section where I provide just enough technical details to help the reader understand the rest of the writeup. Here’s an example background section:

Managed Delivery (MD) supports a GitOps-style workflow. For apps that are on Managed Delivery, engineers can make delivery-related changes to the app by editing a file in their app’s Stash repository called the delivery config.  

To support this workflow, Managed Delivery must be able to identify when a new commit has happened to the default branch of a managed app, and read the delivery config associated with that commit.

The initial implementation of this functionality used a custom Spinnaker pipeline for doing these imports. When an application was onboarded to Managed Delivery, newt would create a special pipeline named import-delivery-config. This pipeline was triggered by commits to the default branch, and would execute a custom pipeline stage that would retrieve the delivery config from Stash and push it to keel, the service that powers Managed Delivery.

This solution, while functional, was inelegant: it exposed an implementation detail of Managed Delivery to end-users, and made it more difficult for users to identify import errors. A better solution would be to have keel identify when commits happen to the repositories of managed apps and import the delivery config directly. This solution was implemented recently, and all apps previously using pipelines were automatically migrated to the native git integration. As will be revealed in the narrative, an unexpected interaction involving the native git integration functionality contributed to this OOPS.

Narrative description

The narrative is the heart of the writeup. If I don’t have enough time to do a complete writeup, then I will just do an executive summary and a narrative description, and skip all of the other sections.

Since the narrative description is often quite long (over ten pages, sometimes many more), I break it up into sections and sub-sections. I typically use the following top-level sections.

  • Prologue
  • The trigger
  • Impact
  • Epilogue


In every OOPS I’ve ever written up, implementation decisions and changes that happen well before the incident play a key role in understanding how the system got into a dangerous state. I use the Prologue section to document these, as well as describing how those decisions were reasonable when they happened.

I break the prologue up into subsections, and I include timeline information in the subsection headers. Here are some examples of prologue subsection headers I’ve used (note: these are from different OOPS writeups).

  • New apps with delivery configs, but aren’t on MD (5 months before impact)
  • Implementing the git integration (4 months before impact)
  • Always using the latest version of a platform library (4 months before impact)
  • A successful <foo> plugin deployment test (8 days before impact)
  • A weekend fix is deployed to staging (4 days before impact)
  • Migrating existing apps (3-4 days before impact)
  • A dependency update eludes dependency locking (1 day before impact)

I often use foreshadowing in my prologue section writeups. Her are some examples:

It will be several months before keel launches its first Titus Run Job orca task. Until one of those new tasks fails, nobody will know that a query against orca for task status can return a payload that keel is incapable of deserializing.

The scope of the query in step 2 above will eventually interact with another part of the system, which will broaden the blast radius of the operational surprise. But that won’t happen for another five months.

Unknown at the time, this PR introduced two bugs:
1. <description of first bug>
2. <description of second bug>
Note that the first bug masks the second. The first bug will become apparent as soon as the code is deployed to production, which will happen in three days. The second bug will lay unnoticed for eleven days.

The trigger

The “trigger” section is the shortest one, but I like to have it as a separate section because it acts as what my colleague J. Paul Reed calls a “pivot point”, a crucial moment in the story of the incident. This section should describe how the system transitions into a state where there is actual customer impact. I usually end the trigger section with some text in red that describes the hazardous state that the system is now in.

Here’s an example of a trigger section:

Trigger: a submitted delivery config

On <date>, at <time>, <name> commits a change to their delivery config that populates the artifacts section. With the delivery config now complete, they submit it to Spinnaker, then point their browser at the environments view of the <app> app, where they can observe Spinnaker manage the app’s deployment.

When <name> submits their delivery config, keel performs the following events:

  1. receives the delivery config via REST API.
  2. deserializes the delivery config from YAML into POJOs.
  3. serializes the config into JSON objects.
  4. writes the JSON objects to the database.

At this point, keel has entered a bad state: it has written JSON objects into the resource table that it will not be able to deserialize. 


The impact section is the longest part of the narrative: it covers everything from the trigger until the system has returned to a stable state. Like the prologue section, I chunk it into subsections. These act as little episodes to make it easier for the reader to follow what’s happening.

Here are examples of some titles for impact subsections I’ve used:

  • User reports access denied on unpin
  • Pinning the library back
  • Maybe it’s gate?
  • Deploying the version with the library pinned back
  • Let’s try rolling back staging
  • Staging is good, let’s do prod
  • Where did the <X> headers go?
  • Rollback to main is complete
  • We’re stable, but why did it break?

For some incidents, I’ll annotate these headers with the timing, like I did in the prologue (e.g., “45 minutes after impact”).

Because so much of our incident coordination is over Slack these days, my impact section will typically have pasted screeenshots of Slack conversation snippets, interspersed with text. I’ll typically write some text that summarizes the interaction, and then paste a screenshot, e.g.:

<name> notes something strange in keel’s it has multiple version parameters where it should only have one:

[Slack screenshot here]

The impact section is mostly written chronologically. However, because it is chunked into episodic subsections, sometimes it’s not strictly in chronological order. I try to emphasize the flow of the narrative over being completely faithful to the ordering of the events. The subsections often describe activities that are going on in parallel, and so describing the incident in the strict ordering of the events would be too difficult to follow.


I’ll usually have an epilogue section that documents work done in the wake of the incident. I split this into subsections as well. An example of a subsection: Fixing the dependency locking issue


Here’s the guidance in the template for the contributors and enablers section:

Various contributors and enablers create vulnerabilities that remain latent in the system (sometimes for long periods of time). Think of these as things that had to be true in order for the incident to take place, or somehow made it worse.

This section is broken up into subsections, one subsection for each contributor. I typically write these at a very low-level of abstraction, where my colleague J. Paul Reed writes these at a higher level.

I think it’s useful to call the various contributors out explicitly because it brings home how complex the incident really was.

Here are some example subsection titles:

  • Violated assumptions about version strings
  • Scope of SQL query
  • Beans not scanned at startup after Titus refactor
  • Incomplete TitusClusterSpecDeserializer
  • Metadata field not populated for PublishedArtifact objects
  • Resilience4J annotations and Kotlin suspend functions
  • Transient errors immediately before deploying to staging
  • Artifact versioning complexity
  • Production pinned for several days
  • No attempts to deploy to production for several days
  • Three large-ish changes landed at about the same time 
  • Holidays and travel
  • Alerts focus on keel errors and resource checks


The guidance we give looks like this:

Which factors helped reduce the impact of this operational surprise?

Like the contributors/enablers section, this is broken up into subsections. Here are some examples of subsection titles:

  • RADAR alerts caught several issues in staging
  • <name> recognized Titus API refactor as a trigger for an issue in production
  • <name> quickly diagnoses artifact metadata issue
  • <name>’s hypothesis about transactions rolling back due to error
  • <name> recognized query too broad
  • <name> notices spike in actuations


Here’s the guidance for this section from the template:

Risks are items (technical architecture or coordination/team related) that created danger in the system. Did the incident reveal any new risks or reinforce the danger of any known risks? (Avoid hindsight bias when describing risks.)

The risks section is where I abstract up some of the contributors to identify higher-level patterns. Here are some example risk subsection titles:

  • Undesired mass actuation
  • Maintaining two similar things in the codebase
  • Problems with dynamic configuration that are only detectable at runtime
  • Plugins that violate assumptions in the main codebase
  • Not deploying to prod for a while

Challenges in handling

Here’s the guidance for this section from the template:

Highlight the obstacles we had to overcome during handling. Was there anything particularly novel, confusing, or otherwise difficult to deal with? How did we figure out what to do? What decisions were made? (Capturing this can be helpful for teaching others how we troubleshoot and improvise). 

In particular, were there unproductive threads of action? Capture avenues that people explored and mitigations that were attempted that did not end up being fruitful.

Sometimes it’s not clear what goes into a contributor and what goes into a challenge. You could put all of these into “contributors” and not write this section at all. However, I think it’s useful to call out what explicitly made the incident difficult to handle. Here are some example subsection headers:

  • Long time to diagnose and remediate
  • Limited signals for making sense of underlying problem
  • Error checking task status as red herring

Other sections

The template has some other sections (incident artifacts, follow-up items, timeline and links), but I often don’t include those in my own writeups. I’ll always do a timeline document as input for writing up the OOPS, and I will typically link it for reference, but I don’t expect anybody to read it. I don’t see the OOPS writeup as the right vehicle for tracking follow-up work, so I don’t put a section in it.

The danger of hidden functional roles

There’s a collection of friends that I have a standing videochat with every couple of weeks. We had been meeting at 8am, but several people developed conflicts at that time, including me. I have a teenager that starts school at 8am, and I’m responsible for getting them to school in the morning (I like to leave the house around 7:40am), which prevented me from participating.

As a group, we decided to reschedule the chat to 7am. This works well for me, because I get up at 6. Today was the first day meeting at the new scheduled time. I got up as I normally do, and was sure to be quiet so as not to wake my wife, Stacy; she sleeps later than I do, but she gets up early enough to rouse our kids for school. I even closed the bedroom door so that any noise I made from the videochat wouldn’t disturb her.

I was on the videochat, taking part in the conversation in hushed tones, when I looked over at the time. I saw it was 7:25am, which is about fifteen minutes before I start getting ready to leave the house. Usually, the rest of the household is up, showering, eating breakfast. But I hadn’t heard a peep from anyone. I went upstairs to discover that nobody else had gotten up yet.

It turns out that my typical morning routine was acting as a natural alarm clock for Stacy. My alarm goes off at 6am every weekday, and I get up, but Stacy stays in bed. However, the noises from my normal morning routine are the thing that rouse her, which is typically around 7am. Today, I was careful to be very quiet, and so she didn’t wake up. I didn’t know that I was functioning as an alarm clock for her! That’s why I was careful to be quiet, and why I didn’t even think to mention to her about the new videochat time.

I suspect this is failure mode is more common than we realize: there is a process inside a system, and over time the process comes to fulfill some unintended, ancillary functional role, and there are people who participate in this process that aren’t even aware of this function.

As an example, consider Chaos Monkey. Chaos Monkey’s intended function is to ensure that engineers design their services to withstand a virtual machine instance failing unexpectedly, by increasing their exposure to this failure mode. But Chaos Monkey also has the unintended effect of recycling instances. For teams that deploy very infrequently, their service might exhibit problems with long-lived instances that they never notice because Chaos Monkey tends to terminate instances before they hit those problems. Now imagine declaring an extended period of time where, in the interest of reducing risk(!), no new code is deployed, and Chaos Monkey is disabled.

When you turn something off, you never know what might break. In some cases, nobody in the system knows.

Plus c’est la même chose, plus ça change

I’m re-reading a David Woods’s paper titled the theory of graceful extensibility: basic rules that govern adaptive systems. The paper proposes a theory to explain how certain types of systems are able to adapt over and over again to changes in their environment. He calls this phenomenon sustained adaptability, which we contrasts with systems that can initially adapt to an environment but later collapse when some feature of the environment changes and they fail to adapt to the new change.

Woods outlines six requirements that any explanatory theory of sustained adaptability must have. Here’s the fourth one (emphasis in the original):

Fourth, a candidate theory needs to provide a positive means for a unit at any scale to adjust how it adapts in the pursuit of improved fitness (how it is well matched to its environment), as changes and challenges continue apace. And this capability must be centered on the limits and perspective of that unit at that scale.

The phrase adjust how it adapts really struck me. Since adaptation is a type of change, this is referring to a second-order change process: these adaptive units have the ability to change the mechanism by which they change themselves! This notion reminded me of Chris Argyris’s idea of double-loop learning.

Woods’s goal is to determine what properties a system must have, what type of architecture it needs, in order to achieve this second-order change process. He outlines in the paper that any such system must be a layered architecture of units that can adapt themselves and coordinate with each other, which he calls a tangled, layered network.

Woods believes there are properties that are fundamental to systems that exhibit sustained adaptability, which implies that these fundamental properties don’t change! A tangled, layered network may reconfigure itself in all sorts of different ways over time, but it must still be tangled and layered (and maintain other properties as well).

The more such systems stay the same, the more they change.

Adapting to a crunch: the Mask Match story

I just got back from Strange Loop, and my favorite talk was Tech When the Sky is Falling: Tools for Crisis Response by Emma Ferguson and Colin Schimmelfing. I’m going to use this talk to illustrate one of the ideas in David Woods theory of graceful extensibility. The idea is that a system needs to deploy, mobilize, or generate capacity when it is at risk of saturation.

My silly doodle of the speakers

Back in March 2020, frontline hospital workers dealing with COVID-19 patients were running short on N95 face masks. Hospitals simply didn’t have enough masks to supply their workers. This shortage of masks is a great example of what Woods calls a crunch, where a system runs short on some resource that it needs. When a system is crunched like this, it needs to adapt. It has to make some sort of change in order to get more of that resource so that it can function properly.

Woods lists three methods for getting more of a resource. If you’ve prepared in advance by stockpiling resources, you can deploy those stockpiles. If you don’t have those extra resources on hand, but your larger network has resources to spare, you can mobilize your network to access those resources. Finally, if you can’t tap into your network to get those resources, your only option is to generate the resources you need. In order to generate resources, you need access to raw materials, and then you need to do work to transform those raw materials into the right resources.

In the case of the mask shortages, the hospitals did not have sufficient stockpiles of N95 masks on hand, so deploying wasn’t an option. It turns out that there were many American households that happened to have N95 masks sitting in storage, and many of those households were willing to donate these unused masks to healthcare workers. In theory, hospitals could mobilize this network of volunteers in order to get these masks to the frontline workers.

There was a problem, though: hospital administrators refused to accept donated N95 masks because of liability concerns. So, this wasn’t something the hospitals were going to do.

Workers wanted masks, and people wanted to donate, but hospital admins wouldn’t let them

Fortunately, there was a loophole: frontline workers could bring in their own masks. Now, the problem to be solved was: how do you get masks from donors who had masks to the workers who wanted them?

Emma and Colin needed to generate a new capability: a mechanism for matching up the donors with the healthcare workers. The raw materials that they initially used to generate this capability were Google Sheets and Gmail for coordinating among the volunteers.

And it worked! However, they quickly ran into a new risk of saturation. Google Sheets has a limit of 50 concurrent editors, and Gmail limits an email account to a maximum of 500 emails per day. And so, once again, the team had to generate a new capability that would scale beyond what Google Sheets and Gmail were capable of. They ended up building a system called Mask Match, by writing a Flask app that they deployed on Heroku, and using Mailgun for sending the emails.

My favorite part of this talk was when Emma Ferguson mentioned that they originally just wanted to pay Google in order to get the Google Sheets and Gmail limits increased (their GoFundMe campaign was quite successful, so getting access to money wasn’t a problem for them). However, they couldn’t figure out how to actually pay Google for a limit increase! This is a wonderful example of what Woods calls brittleness, where a system is unable to extend itself when it reaches its limits. Google is great at building robust systems, but their ethos of removing humans from the loop means that it’s more difficult for consumers of Google services to adapt them to unexpected, emergency scenarios.

The strange beauty of strange loop failure modes

As I’ve posted about previously, at my day job, I work on a project called Managed Delivery. When I first joined the team, I was a little horrified to learn that the service that powers Managed Delivery deploy itself using Managed Delivery.

“How dangerous!”, I thought. What if we push out a change that breaks Managed Delivery? How will we recover? However, after having been on the team for over a year now, I have a newfound appreciation for this approach.

Yes, sometimes there’s something that breaks, and that makes it harder to roll back, because Managed Delivery provides the main functionality for easy rollback. However, it also means that the team gets quite a bit of practice at bypassing Managed Delivery when something goes wrong. They know how to disable Managed Delivery and use the traditional Spinnaker UI to deploy an older version. They know how to poke and prod at the database if the Managed Delivery UI doesn’t respond properly.

These strange loop failure modes are real: if Managed Delivery breaks, we may lose out on the functionality of Managed Delivery to help us recover. But it also means that we’re more ready for handling the situation if something with Managed Delivery goes awry. Yes, Managed Delivery depends on itself, and that’s odd. But we have experience with how to handle things when this strange loop dependency creates a problem. And that is a valuable thing.

Live-drawing my slides during a talk

The other day, I gave an internal talk, and I tried an experiment. Using my iPad and the GoodNotes app, I drew all of my slides while I was talking (except the first slide, which I drew in advance).

“What font is that?” someone asked. It’s my handwriting

I’ve always been in awe of people who can draw, I’ve never been good at it.

“Where’s the bug”, it says. Not my best handwriting

Over the years, I’ve tried doodling more. I was influenced by Dan Roam’s books, Julia Evans’s zines, sketchnotes, and most recently, Christina Wodtke’s Pencil Me In.

The words have stink lines, so you know they’re bad

If you’ve read my blog before, you’ve seen some of my previous doodles (e.g., Root cause of failure, root cause of success or Taming complexity: from contract to compact).

We need to complete the action items so it never happens again

When I was asked to present to a team, I wanted to use my drawings rather than do traditional slides. I actually hate using tools like PowerPoint and Google Slides to do presentations. Typically I use Deckset, but in this case, I wanted to do them all drawn.

A different perspective on incidents

I started off by drawing out my slides in advance. But then I thought, “instead of showing pre-drawn slides, why don’t I draw the slides as I talk? That way, people will know where to look because they’ll look at where I’m drawing.”

I still had to prepare the presentation in advance. I drew all of the slides beforehand. And then I printed them out and had them in front of me so that I could re-draw them during the talk. Since it was done over Zoom, people couldn’t actually see that I was working from the print-outs (although they might have heard the paper rustling).

Contributing factors aren’t like root cause

One benefit of this technique was that it made it easier to answer questions, because I could draw out my answer. When I was writing the text at the top, somebody asked, “Is that something like a root cause chain?” I drew the boxes and arrows in response, to explain how this isn’t chain-like, but instead is more like a web.

The selected images above should give you a sense of what my slides looked like. I had fun doing the presentation, and I’d try this approach again. It was certainly more enjoyable than futzing with slide layout.

Useful knowledge and improvisation

Eric Dobbs recently retold a story on twitter (a copy is on his wiki) about one of his former New Relic colleagues, Nicholas Valler.

At the time, Nicholas was new to the company. He had just discovered a security vulnerability, and then (unrelated to that security vulnerability), an incident happened and, well, I encourage you to read the whole story first, and then come back to this post.

In the end, the engineers were able to leverage the security vulnerability to help resolve the incident. As is my wont, I made a snarky comment.

But I did want to make a more serious comment about what this story illustrates. In a narrow sense, this security vulnerability helped the New Relic engineers remediate before there was severe impact. But in a broader sense, the following aspects helped them remediate:

  • they had useful knowledge of some aspect of the the system (port 22 was open to the world)
  • they could leverage that knowledge to improvise a solution (they could use this security hole to log in and make changes to the kafka configuration)

The irony here is that it was a new employee that had the useful knowledge. Typically, it’s the tenured engineers who have this sort of knowledge, as they’ve accumulated it with experience. In this case, the engineer discovered this knowledge right before it was needed. That’s what make this such a great story!

I do think that how Nicholas found it, by “poking around”, is a behavior that comes with general experience, even though he didn’t have much experience at the company.

But being in possession of useful knowledge isn’t enough. You also need to be able to recognize when the knowledge is useful and bring it to bear.

These two attributes: having useful knowledge about the system and the ability to apply that knowledge to improvise a solution, are critical for being able to deal effectively with incidents. Applying these are resilience in action.

It’s not a focus of this particular story, but, in general, this sort of knowledge is distributed across individuals. This means that it’s the ad-hoc team that forms during an incident that needs to possess these attributes.


Back in July, Ray Ashman at Mailchimp posted a wonderful writeup of an internal incident (h/t to SRE Weekly). It took the Mailchimp engineers almost two days to make sense of the failure mode.

The trigger was a change to a logging statement, in order to log an exception. During the incident, the engineers noticed that this change lined up with the time that the alerts fired. But, other than the timing, there wasn’t any evidence to suggest that the log statement change was problematic. The change didn’t have any apparent relationship to the symptoms they were seeing with the job runner, which was in a different part of the codebase. And so they assumed that the logging statement change was innocuous.

As it happened, there was a coupling between that log statement and the job runner. Unfortunately for the engineers, this coupling was effectively invisible to them. The connection between the logging statement and the job running was Mailchimp’s log processing pipeline. Here’s an excerpt from the writeup (emphasis mine):

Our log processing pipeline does a bit of normalization to ensure that logs are formatted consistently; a quirk of this processing code meant that trying to log a PHP object that is Iterable would result in that object’s iterator methods being invoked (for example, to normalize the log format of an Array).

Normally, this is an innocuous behavior—but in our case, the harmless logging change that had shipped at the start of the incident was attempting to log PHP exception objects. Since they were occurring during job execution, these exceptions held a stacktrace that included the method the job runner uses to claim jobs for execution (“locking”)—meaning that each time one of these exceptions made it into the logs, the logging pipeline itself was invoking the job runner’s methods and locking jobs that would never be actually run! 

Fortunately, there were engineers who had experience with this failure mode before:

Since the whole company had visibility into our progress on the incident, a couple of engineers who had been observing realized that they’d seen this exact kind of issue some years before.

Having identified the cause, we quickly reverted the not-so-harmless logging change, and our systems very quickly returned to normal.

In the moment, the engineers could not conceive of how a change in behavior in the job runner could be affected by the modification of a log statement in an unrelated part of the code. It was literally unthinkable to them.