Choosing Good SLIs

Transitioning from an on-prem environment to a cloud environment involves a lot of major shifts for organizations. One of those shifts is often around how we monitor the overall health of systems. The typical way to measure things like the availability, reliability, and performance of systems is with SLIs or Service Level Indicators. SLIs are a valuable tool both on-prem and in the cloud, but when it comes to the latter, I often see organizations carrying over some operational anti-patterns from their data center environment.

Unlike public clouds, data centers are often resource-constrained. Services run on dedicated sets of VMs and it can take days or weeks for new physical servers to be provisioned. Consequently, it’s common for organizations to closely monitor metrics such as CPU utilization, memory consumption, disk space, and so forth since these are all precious resources within a data center.

Often what happens is that ops teams get really good at identifying and pattern-matching the common issues that arise in their on-prem environment. For instance, certain applications may be prone to latency issues. Each time we dig into a latency issue we find that the problem is due to excessive garbage collection pauses. As a result, we define a metric around garbage collection because it is often an indicator of performance problems in the application. In practice, this becomes an SLI, whether it’s explicitly defined as such or not, because there is some sort of threshold beyond which garbage collection is considered “excessive.” We begin watching this metric closely to gauge whether the service is healthy or not and alerting on it.

The cloud is a very different environment than on-prem. Whether we’re using an orchestrator such as Kubernetes or a serverless platform, containers are usually ephemeral and instances autoscale up and down. If an instance runs out of memory, it will just get recycled. This is why we sometimes say you can “pay your way out” of a problem in these environments because autoscaling and autohealing can hide a lot of application issues such as a slow memory leak. In an on-prem environment, these can be significantly more impactful. The performance profile of applications often looks quite differently in the cloud than on-prem as well. Underlying hardware, tenancy, and networking characteristics differ considerably. All this is to say, things look and behave quite differently between the two environments, so it’s important to reevaluate operational practices as well. With SLIs and monitoring, it’s easy to bias toward specific indicators from on-prem, but they might not translate to more cloud-native environments.

User-centric monitoring

So how do we choose good SLIs? The key question to ask is: what is the customer’s experience like? Everything should be driven from this. Is the application responding slowly? Is it returning errors to the user? Is it returning bad or incorrect results? These are all things that directly impact the customer’s experience. Conversely, things that do not directly impact the customer’s experience are questions such as what is the CPU utilization of the service? The memory consumption? The rate of garbage collection cycles? These are all things that could impact the customer’s experience, but without actually looking from the user’s perspective, we simply don’t know whether they are or not. Rather, they are diagnostic tools that—once an issue is identified—can help us to better understand the underlying cause.

Take, for example, the CPU and memory utilization of processes on your computer. Most people probably are not constantly watching the Activity Monitor on their MacBook. Instead, they might open it up when they notice their machine is responding slowly to see what might be causing the slowness.

Three key metrics

When it comes to monitoring services, there are really three key metrics that matter: traffic rate, error rate, and latency. These three things all directly impact the user’s experience.

Traffic Rate

Traffic rate, which is usually measured in requests or queries per second (qps), is important because it tells us if something is wrong upstream of us. For instance, our service might not be throwing any errors, but if it’s suddenly handling 0 qps when it ordinarily is handling 80-100 qps, then something happened upstream that we should know about. Perhaps there is a misconfiguration that is preventing traffic from reaching our service, which almost certainly impacts the user experience.

Traffic rate or qps for a service

Error Rate

Error rate simply tells us the rate in which the service is returning errors to the client. If our service normally returns 200 responses but suddenly starts returning 500 errors, we know something is wrong. This requires good status code hygiene to be effective. I’ve encountered codebases where various types of error codes are used to indicate non-error conditions which can add a lot of noise to this type of SLI. Additionally, this metric might be more fine-grained than just “error” or “not error”, since—depending on the application—we might care about the rate of specific 2xx, 4xx, or 5xx responses, for example.

It’s common for teams to rely on certain error logs rather than response status codes for monitoring. This can provide even more granularity around types of error conditions, but in my experience, it usually works better to rely on fairly coarse-grained signals such as HTTP status codes for the purposes of aggregate monitoring and SLIs. Instead, use this logging for diagnostics and troubleshooting once you have identified there is a problem (I am, however, a fan of structured logging and log-based metrics for instrumentation but this is for another blog post).

Response codes for a service

Latency

Combined with error rate, latency tells us what the customer’s experience is really like. This is an important metric for synchronous, user-facing APIs but might be less critical for asynchronous processes such as services that consume events from a message queue. It’s important to point out that when looking at latency, you cannot use averages. This is a common trap I see ops teams and engineers fall into. Latency rarely follows a normal distribution, so relying on averages or medians to provide a summarized view of how a system is performing is folly.

Instead, we have to look at percentiles to get a better understanding of what the latency distribution looks like. Similarly, you cannot average percentiles either. It mathematically makes no sense, meaning you can’t, for instance, look at the average 90th percentile over some period of time. To summarize latency, we can plot multiple percentiles on a graph. Alternatively, heatmaps can be an effective way to visualize latency because they can reveal useful details like distribution modes and outliers. For example, the heatmap below shows that the latency for this service is actually bimodal. Requests usually either respond in approximately 10 milliseconds or 1 second. This modality is not apparent in the line chart above the heatmap where we are only plotting the 50th, 95th, and 99th percentiles. The line chart does, however, show that latency ticked up a tiny bit around 10:10 AM following a severe spike in tail latency where the 99th percentile momentarily jumped over 4 seconds…curious.

Latency distribution for a service as percentiles
Latency distribution for a service as a heatmap

Identifying other SLIs

While these three metrics are what I consider the critical baseline metrics, there may be other SLIs that are important to a service. For example, if our service is a cache, we might care about the freshness of data we’re serving as something that impacts the customer experience. If our service is queue-based, we might care about the time messages spend sitting in the queue.

Heatmap showing the age distribution of data retrieved from a cache

Whatever the SLIs are, they should be things that directly matter to the user’s experience. If they aren’t, then at best they are a useful diagnostic or debugging tool and at worst they are just dashboard window dressing. Usually, though, they’re no use for proactive monitoring because it’s too much noise, and they’re no use for reactive debugging because it’s typically pre-aggregated data.

What’s worse is that when we focus on the wrong SLIs, it can lead us to take steps that actively harm the customer’s experience or simply waste our own time. A real-world example of this is when I saw a team that was actively monitoring garbage collection time for a service. They noticed one instance in particular appeared to be running more garbage collections than the others. While it appeared there were no obvious indicators of latency issues, timeouts, or out-of-memory errors that would actually impact the client, the team decided to redeploy the service in order to force instances to be recycled. This redeploy ended up having a much greater impact on the user experience than any of the garbage collection behavior ever did. The team also spent a considerable amount of time tuning various JVM parameters and other runtime settings, which ultimately had minimal impact.

Where lower-level metrics can provide value is with optimizing resource utilization and cloud spend. While the elastic nature of cloud may allow us to pay our way out of certain types of problems such as a memory leak, this can lead to inefficiency and waste long term. If we see that our service only utilizes 20% of its allocated CPU, we are likely overprovisioned and could save money. If we notice memory consumption consistently creeping up and up before hitting an out-of-memory error, we likely have a memory leak. However, it’s important to understand this distinction in use cases: SLIs are about gauging customer experience while these system metrics are for identifying optimizations and understanding long-term resource characteristics of your system. At any rate, I think it’s preferable to get a system to production with good monitoring in place, put real traffic on it, and then start to fine-tune its performance and resource utilization versus trying to optimize it beforehand through synthetic means.

Transitioning from an on-prem environment to the cloud necessitates a shift in how we monitor the health of systems. It’s essential to recognize and discard operational anti-patterns from traditional data center environments, where resource constraints often lead to a focus on specific metrics and behaviors. This can frequently lead to a sort of “overfitting” when monitoring cloud-based systems. The key to choosing good SLIs is by aligning them with the customer’s experience. Metrics such as traffic rate, error rate, and latency directly impact the user and provide meaningful insights into the health of services. By emphasizing these critical baseline metrics and avoiding distractions with irrelevant indicators, organizations can proactively monitor and improve the customer experience. Focusing on the right SLIs ensures that efforts are directed toward resolving actual issues that matter to users, avoiding pitfalls that can inadvertently harm user experience or waste valuable time. As organizations navigate the complexities of migrating to a cloud-native environment, a user-centric approach to monitoring remains fundamental to successful and efficient operations.

Need help making the transition?

Real Kinetic helps organizations with their cloud migrations and implementing effective operations. If you have questions or need help getting started, we’d love to hear from you. These emails come directly to us, and we respond to every one.

Cloud without Kubernetes

I think it’s safe to say Kubernetes has “won” the cloud mindshare game. If you look at the CNCF Cloud Native landscape (and manage to not go cross eyed), it seems like most of the projects are somehow related to Kubernetes. KubeCon is one of the fastest-growing industry events. Companies we talk to at Real Kinetic who are either preparing for or currently executing migrations to the cloud are centering their strategies around Kubernetes. Those already in the cloud are investing heavily in platform-izing their Kubernetes environment. Kubernetes competitors like Nomad, Pivotal Cloud Foundry, OpenShift, and Rancher have sort of just faded to the background (or simply pivoted to Kubernetes). In many ways, “cloud native” seems to be equated with “Kubernetes”.

All this is to say, the industry has coalesced around Kubernetes as the way to do cloud. But after working with enough companies doing cloud, watching their experiences, and understanding their business problems, I can’t help but wonder: should it be? Or rather, is Kubernetes actually the right level of abstraction?

Going k8sless

While we’ve worked with a lot of companies doing Kubernetes, we’ve also worked with some that are deliberately not. Instead, they leaned into serverless—heavily—or as I like to call it, they’ve gone k8sless. These are not small companies or startups, they are name brands you would recognize.

At first, we were skeptical. Our team came from a company that made it all the way to IPO using Google App Engine, one of the earliest serverless platforms available. We have regularly espoused the benefits of serverless. We’ve talked to clients about how they should consider it for their own workloads (often to great skepticism). But using only serverless? For once, we were the serverless skeptics. One client in particular was beginning a migration of their e-commerce platform to Google Cloud. They wanted to do it completely serverless. We gave our feedback and recommendations based on similar migrations we’ve performed:

“There are workloads that aren’t a good fit.”

“It would require major re-architecting.”

“It will be expensive once fully migrated.”

“You’ll have better cost efficiency bin packing lots of services into VMs with Kubernetes.”

We articulated all the usual arguments made by the serverless doubters. Even Google was skeptical, echoing our sentiments to the customer. “Serious companies doing online retail like The Home Depot or Target are using Google Kubernetes Engine,” was more or less the message. We have a team of serverless experts at Real Kinetic though, so we forged ahead and helped execute the migration.

Fast forward nearly three years later and we will happily admit it: we were wrong. You can run a multibillion-dollar e-commerce platform without a single VM. You don’t have to do a full rewrite or major re-architecting. It can be cost-effective. It doesn’t require proprietary APIs or constraints that result in vendor lock-in. It might sound like an exaggeration, but it’s not.

Container as the interface

Over the last several years, Google’s serverless offerings have evolved far beyond App Engine. It has reached the point where it’s now viable to run a wide variety of workloads without much issue. In particular, Cloud Run offers many of the same benefits of a PaaS like App Engine without the constraints. If your code can run in a container, there’s a very good chance it will run on Cloud Run with little to no modification.

In fact, other than using the gcloud CLI to deploy a service, there’s nothing really Google- or Cloud Run-specific needed to get a functioning application. This is because Cloud Run uses Knative, an open-source Kubernetes-based platform, as its deployment interface. And while Cloud Run is a Google-managed backend for the Knative interface, we could just as well switch the backend to GKE or our own Kubernetes cluster. When we implement our Cloud Run services, we actually implement them using a Kubernetes Deployment manifest, shown below, and right before deploying, we swap Deployment for Knative’s Service manifest.

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    cloud.googleapis.com/location: us-central1
    service: my-service
  name: my-service
spec:
  template:
    spec:
      containers:
        - image: us.gcr.io/my-project/my-service:v1
          name: my-service
          ports:
            - containerPort: 8080
          resources:
            limits:
              cpu: 2
              memory: 1024Mi

This means we can deploy to Kubernetes without Knative at all, which we often do during development using the combination of Skaffold and K3s to perform local testing. It also allows us to use Kubernetes native tooling such as Kustomize to manage configuration. Think of Cloud Run as a Kubernetes Deployment as a service (though really more like Deployment and Service…as a service).

“Normal” businesses versus internet-scale businesses

What about cost? Yes, the unit cost in terms of compute is higher with serverless. If you execute enough CPU cycles to fill the capacity of a VM, you are better off renting the whole VM as opposed to effectively renting timeshares of it. But here’s the thing: most “normal” businesses tend to have highly cyclical traffic patterns throughout the day and their scale is generally modest.

What do I mean by “normal” businesses? These are primarily non-internet-scale companies such as insurance, fast food, car rental, construction, or financial services, not Google, Netflix, or Amazon. As a result, these companies can benefit greatly from pay-per-use, and those in the retail space also benefit greatly from the elasticity of this model during periods like Black Friday or promotional campaigns. Businesses with brick-and-mortar have traffic that generally follows their operating hours. During off-hours, they can often scale quite literally to zero.

Many of these businesses, for better or worse, treat software development as an IT cost center to be managed. They don’t need—or for that matter, want—the costs and overheads associated with platform-izing Kubernetes. A lot of the companies we interact with fall into this category of “normal” businesses, and I suspect most companies outside of tech do as well.

BYOP—Bring Your Own Platform

I’ve asked it before: is Kubernetes really the end-game abstraction? In my opinion, it’s an implementation detail. I don’t think I’m alone in that opinion. Some companies put a tremendous amount of investment into abstracting Kubernetes from their developers. This is what I mean by “platform-izing” Kubernetes. It typically involves significant and ongoing OpEx investment. The industry has started to coalesce around two concepts that encapsulate this: Platform Engineering and Internal Developer Platform. So while Kubernetes may have become the default container orchestrator, the higher-level pieces—the pieces constituting the Internal Developer Platform—are still very much bespoke. Kelsey Hightower said it best: the majority of people managing infrastructure just want a PaaS. The only requirement: it has to be built by them. That’s a problem.

Imagine having a Kubernetes cluster per Deployment. Full blast radius isolation, complete cost traceability, granular yet simple permissioning. It sounds like a maintenance nightmare though, right? Now imagine those clusters just being hidden from you completely and the Deployment itself is the only thing you interact with and maintain. You just provide your container (or group of containers), configure your CPU and memory requirements, specify the network and resource access, and deploy it. The Deployment manages your load balancing and ingress, automatically scales the pods up and down or canaries traffic, and gives you aggregated logs and metrics out of the box. You only pay for the resources consumed while processing a request. Just a few years ago, this was a futuristic-sounding fantasy.

The platform Kelsey describes above does now exist. From my experience, it’s a nearly ideal solution for those “normal” businesses who are looking to minimize complexity and operational costs and avoid having to bring (more like build) their own platform. I realize GCP is a distant third when it comes to public cloud market share so this will largely fall on deaf ears, but for those who are still listening: stop wasting time on Kubernetes and just use Cloud Run. Let me expand on the reasons why.

  1. Easily and quickly get started with the cloud. Many of the companies we work with who are still in the midst of migrating to the cloud get hung up with analysis paralysis. Cloud Run isn’t a perfect solution for everything, but it’s good enough for the majority of cases. The rest can be handled as exceptions.

  2. Minimize complexity of cloud environments. Cloud Run does not eliminate the need for infrastructure (there are still caches, queues, databases, and so forth), but it greatly simplifies it. Using managed services for the remaining infrastructure pieces simplifies it further.

  3. Increase the efficiency of your developers and reduce operational costs. Rather than spending most of their time dealing with infrastructure concerns, allow your developers to focus on delivering business value. For most businesses, infrastructure is undifferentiated commodity work. By “outsourcing” large parts of your undifferentiated Internal Developer Platform, you can reallocate developers to product or feature development and reduce operational costs. This allows you to get the benefits of Platform Engineering with a fraction of the maintenance and overhead. Lastly, if you are a “normal” business that doesn’t operate at internet scale and has fairly cyclical traffic, it’s entirely likely Cloud Run will be cheaper than VM-based platforms.

  4. Maintain the flexibility to evolve to a more complex solution over time if needed. This is where traditional serverless platforms and PaaS solutions fall short. Again, with Cloud Run there is no actual vendor lock-in, it’s just a Kubernetes Deployment as a Service. Even without Knative, we can take that Deployment and run it in any Kubernetes cluster. This is a very different paradigm from, say, App Engine where you wrote your application using App Engine APIs and deployed your service to the App Engine runtime. In this new paradigm, the artifact is a Plain Old Container. There are cases where Cloud Run is not a good fit, such as certain kinds of stateful legacy applications or services with sustained, non-cyclical traffic. We don’t want to be painted into a corner with these types of situations so having flexibility is important.

There are similar analogs to Cloud Run on other cloud platforms. For example, AWS has AppRunner. However, in my experience these fall short in terms of developer experience because of either lack of investment from the cloud provider or environment complexity (as I would argue is the case for AWS). Managed services like Cloud Run are one of the areas that GCP truly excels and differentiates itself.

Just use Cloud Run, seriously

I realize not everyone will be convinced. The gravitational pull of Kubernetes is strong and as a platform, it’s a safe bet. However, operationalizing Kubernetes properly—whether it’s a managed offering like GKE or not—requires some kind of platform team and ongoing investment. We’ve seen it approached without this where developers are given clusters or allowed to spin them up and fend for themselves. This quickly becomes untenable because standards are non-existent, security and compliance is unmanageable, and developer time is split between managing infrastructure and actual feature development.

If your organization is unable or unwilling to make this investment, I urge you to consider Cloud Run. There’s still work needed on the periphery to properly operationalize it, such as implementing CI/CD pipelines and managing accessory infrastructure, but it’s a much lower investment. Additionally, it provides an escape hatch—unlike App Engine or traditional PaaS solutions, there is no real switching cost in moving to Kubernetes if you need to in the future. With Cloud Run, serverless has finally reached a tipping point where it’s now viable for a majority of workloads rather than a niche subset. Unlike Kubernetes, it provides the right level of abstraction for most businesses building software. In my opinion, serverless is still not taken seriously due to preconceived notions, but it’s time to start reevaluating those notions.

Agree? Disagree? I’d love to hear your thoughts. If you’re an organization that would like to do cloud differently or are looking for the playbook to operationalize Google Cloud Platform, please get in touch.

Meeting notes lose value the moment you finish writing them—and it’s time to fix that

I like to be prepared in meetings. In some ways it’s probably an innate part of my personality, but it also became more important to me as my role has changed throughout my career. In particular, the first time I became an engineering manager is when I started to become a more diligent notetaker and meeting preparer. I think this is largely because my job shifted from being output-centric to more people- and meeting-centric. I still took notes and prepared when I was a software engineer, but it was for a very different context and purpose. As an engineer, my work centered around code output. As a manager, my work instead centered around coordinating, following up, and supporting my team. If you’ve never worked as a manager before, this probably just sounds like paper-pushing, but it’s actually a lot of work—and important! The work product is just different from that of an individual contributor.

When I became a manager, I began taking meeting notes in a small Moleskine notebook. For every meeting, I’d write down the meeting name and the date. I would try to jot down salient points or context, questions, things I wanted to follow up on, or action items I needed to do or delegate. As you can see below, it’s messy. Really messy. It never felt like a particularly good solution. It was hard to find things, hard to pluck out the important action items or follow-ups, hard to even remember who was in a meeting without cross-referencing my calendar. Not to mention my terrible handwriting meant even just reading my own notes was difficult.

An interesting thing about the human brain is that it’s inherently selfish—that is, it’s really good at remembering things that are important to us. The things that are top of mind are probably not things I need to actually write down to remember. But most managers are likely getting pulled in a lot of different directions with a lot of different asks that are all competing for those limited brain cycles. Really good managers seem to have a special knack for juggling all of these things. It’s also why you often hear managers talk about how tiring their job is even though it seems like all they do is go to meetings!

The hard truth about my note-taking system is that I would take a lot of notes, write a lot of action items, and feel really productive in my meetings. Then I would proceed to never look at those notes again. Partly because of the chaos of meeting-packed days week after week, but also because it’s just hard to derive value from notes. Countless times a topic or question would come up in a discussion where I knew I had notes from a previous meeting about it, but it was just impossible to actually find anything in a notebook full of hastily scribbled notes. And by the time you find it, the conversation has moved on. You know how people say your new car loses its value the moment you drive off the lot? Your meeting notes lose value the moment you finish writing them.

This leads to another interesting thing about the human brain—it’s pretty good at organizing memories around time and people. “I remember talking to Joe about managing our cloud costs last week in our weekly cloud strategy meeting”—that sort of thing. And while my notebook provided a chronological ordering of my meeting notes, it wasn’t really conducive to recalling important information quickly or managing my to-do list.

A software engineer’s job often involves coordinating across different software systems, but their to-do list likely consists of things along the lines of “do X.” This is why tools like Jira or Asana exist, to manage the backlog of X’s that need to be done and provide visibility for the people coordinating those X’s.

A manager’s job involves coordinating across a different kind of system—people. A manager’s to-do list is going to consist mostly of things like “talk to Y about Z.” Again, the work product is different. It’s about making sure there is alignment and lines of communication between various people or teams. Your work shifts from being a do-er to a delegate-er and communicator. This kind of work is not managed in Jira or displayed in a Gantt chart. It’s often not managed anywhere except perhaps scribbled in the depths of a Moleskine notebook or tucked away in the corner of a meeting-fatigued brain.

Nevertheless, I carried on with my note-taking system of questionable value, even after transitioning back to an individual contributor role. It wasn’t until I started consulting that I had a realization. With consulting, I work with a lot of different people across a lot of different projects across a lot of different clients. The type of consulting we do at Real Kinetic is very discussional in nature. While we have deliverables, most of our work product is in the form of discussion, guidance, recommendations, coaching, and helping organizations with their own communication challenges. It’s not work that can be managed in a traditional task-management system. Instead, it’s much like the manager’s work of connecting threads of conversation across meetings and people and juggling lots of asks from clients.

For example, in a meeting with John I might realize we need to connect with Rachel to talk about strategies for improving development velocity. Sure, you could maybe put “Talk to Rachel about dev velocity” into a Trello card or a to-do list app but in doing so it loses the surrounding context. And for a role that is more discussion-oriented than task-oriented, the context is important. Not only that, but tools like Trello or Todoist are just not really designed for this purpose. They are meant more for the do-ers, not the delegate-ers or communicators. They are clunky to use for someone whose job consists mostly of being in meetings and talking to people day in and day out. This is the challenge with productivity apps—most of them are centered around task management and task collaboration. And actual note-taking apps like Evernote are definitely not designed to solve this because they are intended to replace my Moleskine notebook filled with notes I will never look at again.

Now, coming back to my realization: I realized that my meeting notes were not valuable in and of themselves. Rather, they were the medium for my meeting-centric work management. Unfortunately, my notebook was not a great solution, nor was Evernote, nor Google Docs.

What I was really looking for was a sort of to-do list oriented around people and meetings and driven from my meeting notes. Not something centered around task management or collaboration or notes as being anything other than incidental to the process. Instead, I was looking for a tool that could synthesize my notes into something valuable and actionable for me. And I never found it, which is why we ended up creating Witful.

The idea behind Witful is a productivity app for the people whose jobs revolve around, well, people. It turns your meeting notes into something much more valuable. Now, I can take my meeting notes similar to how I used to, but rather than important items falling by the wayside, those items are surfaced to me. Witful tells me if I need to prepare for an upcoming meeting, if there are takeaways from a meeting I need to follow up on, or action items I need to address. And much like the way our brain organizes information, Witful indexes all of my meeting-related content around my meetings, the people in those meetings, and time, making it easy to quickly recall information.

Witful has not radically altered the way I approach meetings. Instead, what it’s done is augmented my previous workflow. It gives me a central place for all my meeting notes, much like the Moleskine notebook did, except it lets me extract much more value from those notes. This has helped me with my consulting work because it has given me the same uncanny knack for juggling lots of things that those really good managers I’ve worked with seem to have. If you’re not a meeting note-taker, Witful might not be for you. If you are and your current system has never felt quite right, you’d like to get more value out of your notes, or you’re looking for a meeting-centric work management system, you should give it a shot.

SRE Doesn’t Scale

We encounter a lot of organizations talking about or attempting to implement SRE as part of our consulting at Real Kinetic. We’ve even discussed and debated ourselves, ad nauseam, how we can apply it at our own product company, Witful. There’s a brief, unassuming section in the SRE book tucked away towards the tail end of chapter 32, “The Evolving SRE Engagement Model.” Between the SLIs and SLOs, the error budgets, alerting, and strategies for handling change management, it’s probably one of the most overlooked parts of the book. It’s also, in my opinion, one of the most important.

Chapter 32 starts by discussing the “classic” SRE model and then, towards the end, how Google has been evolving beyond this model. “External Factors Affecting SRE”, under the “Evolving Services Development: Frameworks and SRE Platform” heading, is the section I’m referring to specifically. This part of the book details challenges and approaches for scaling the SRE model described in the preceding chapters. This section describes Google’s own shift towards the industry trend of microservices, the difficulties that have resulted, and what it means for SRE. Google implements a robust site reliability program which employs a small army of SREs who support some of the company’s most critical systems and engage with engineering teams to improve the reliability of their products and services. The model described in the book has proven to be highly effective for Google but is also quite resource-intensive. Microservices only serve to multiply this problem. The organizations we see attempting to adopt microservices along with SRE, particularly those who are doing it as a part of a move to cloud, frequently underestimate just how much it’s about to ruin their day in terms of thinking about software development and operations.

It is not going from a monolith to a handful of microservices. It ends up being hundreds of services or more, even for the smaller companies. This happens every single time. And that move to microservices—in combination with cloud—unleashes a whole new level of autonomy and empowerment for developers who, often coming from a more restrictive ops-controlled environment on prem, introduce all sorts of new programming languages, compute platforms, databases, and other technologies. The move to microservices and cloud is nothing short of a Cambrian Explosion for just about every organization that attempts it. I have never seen this not play out to some degree, and it tends to be highly disruptive. Some groups handle it well—others do not. Usually, however, this brings an organization’s delivery to a grinding halt as they try to get a handle on the situation. In some cases, I’ve seen it take a year or more for a company to actually start delivering products in the cloud after declaring they are “all in” on it. And that’s just the process of starting to deliver, not actually delivering them.

How does this relate to SRE? In the book, Google says a result of moving towards microservices is that both the number of requests for SRE support and the cardinality of services to support have increased dramatically. Because each service has a base fixed operational cost, even simple services demand more staffing. Additionally, microservices almost always imply an expectation of lower lead time for deployment. This is invariably one of the reasons we see organizations adopting them in the first place. This reduced lead time was not possible with the Production Readiness Review model they describe earlier in chapter 32 because it had a lead time of months. For many of the organizations we work with, a lead time of months to deliver new products and capabilities to their customers is simply not viable. It would be like rewinding the clock to when they were still operating on prem and completely defeat the purpose of microservices and cloud.

But here’s the key excerpt from the book: “Hiring experienced, qualified SREs is difficult and costly. Despite enormous effort from the recruiting organization, there are never enough SREs to support all the services that need their expertise.” The authors conclude, “the SRE organization is responsible for serving the needs of the large and growing number of development teams that do not already enjoy direct SRE support. This mandate calls for extending the SRE support model far beyond the original concept and engagement model.”

Even Google, who has infinite money and an endless recruiting pipeline, says the SRE model—as it is often described by the people we encounter referencing the book—does not scale with microservices. Instead, they go on to describe a more tractable, framework-oriented model to address this through things like codified best practices, reusable solutions, standardization of tools and patterns, and, more generally, what I describe as the “productization” of infrastructure and operations.

Google enforces standards and opinions around things like programming languages, instrumentation and metrics, logging, and control systems surrounding traffic and load management. The alternative to this is the Cambrian Explosion I described earlier. The authors enumerate the benefits of this approach such as significantly lower operational overhead, universal support by design, faster and lower overhead SRE engagements, and a new engagement model based on shared responsibility rather than either full SRE support or no SRE support. As the authors put it, “This model represents a significant departure from the way service management was originally conceived in two major ways: it entails a new relationship model for the interaction between SRE and development teams, and a new staffing model for SRE-supported service management.”

For some reason, this little detail gets lost and, consequently, we see groups attempting to throw people at the problem, such as embedding an SRE on each team. In practice, this usually means two things: 1) hiring a whole bunch of SREs—which even Google admits to being difficult and costly—and 2) this person typically just becomes the “whipping boy” for the team. More often than not, this individual is some poor ops person who gets labeled “SRE.”

With microservices, which again almost always hit you with a near-exponential growth rate once you adopt them, you simply cannot expect to have a handful of individuals who are tasked with understanding the entirety of a microservice-based platform and be responsible for it. SRE does not mean developers get to just go back to thinking about code and features. Microservices necessitate developers having skin in the game, and even Google has talked about the challenges of scaling a traditional SRE model and why a different tack is needed.

“The constant growth in the number of services at Google means that most of these services can neither warrant SRE engagement nor be maintained by SREs. Regardless, services that don’t receive full SRE support can be built to use production features that are developed and maintained by SREs. This practice effectively breaks the SRE staffing barrier. Enabling SRE-supported production standards and tools for all teams improves the overall service quality across Google.”

My advice is to stop thinking about SRE as an implementation specifically and instead think about the problems it’s solving a bit more abstractly. It’s unlikely your organization has Google-level resources, so you need to consider the constraints. You need to think about the roles and responsibilities of developers as well as your ops folks. They will change significantly with microservices and cloud out of necessity. You’ll need to think about how to scale DevOps within your organization and, as part of that, what “DevOps” actually means to your organization. In fact, many groups are probably better off simply removing “SRE” and “DevOps” from their vocabulary altogether because they often end up being distracting buzzwords. For most mid-to-large-sized companies, some sort of framework- and platform- oriented model is usually needed, similar to what Google describes.

I’ve seen it over and over. This hits companies like a ton of bricks. It requires looking at some hard org problems. A lot of self-reflection that many companies find uncomfortable or just difficult to do. But it has to be done. It’s also an important piece of context when applying the SRE book. Don’t skip over chapter 32. It might just be the most important part of the book.


Real Kinetic helps clients build great engineering organizations. Learn more about working with us.

Structuring a Cloud Infrastructure Organization

Real Kinetic often works with companies just beginning their cloud journey. Many come from a conventional on-prem IT organization, which typically looks like separate development and IT operations groups. One of the main challenges we help these clients with is how to structure their engineering organizations effectively as they make this transition. While we approach this problem holistically, it can generally be looked at as two components: product development and infrastructure. One might wonder if this is still the case with the shift to DevOps and cloud, but as we’ll see, these two groups still play important and distinct roles.

We help clients understand and embrace the notion of a product mindset as it relates to software development. This is a fundamental shift from how many of these companies have traditionally developed software, in which development was viewed as an IT partner beholden to the business. This transformation is something I’ve discussed at length and will not be the subject of this conversation. Rather, I want to spend some time talking about the other side of the coin: operations.

Operations in the Cloud

While I’ve talked about operations in the context of cloud before, it’s only been in broad strokes and not from a concrete, organizational perspective. Those discussions don’t really get to the heart of the matter and the question that so many IT leaders ask: what does an operations organization look like in the cloud?

This, of course, is a highly subjective question to which there is no “right” answer. This is doubly so considering that every company and culture is different. I can only humbly offer my opinion and answer with what I’ve seen work in the context of particular companies with particular cultures. Bear this in mind as you think about your own company. More often than not, the cultural transformation is more arduous than the technology transformation.

I should also caveat that—outside of being a strategic instrument—Real Kinetic is not in the business of simply helping companies lift-and-shift to the cloud. When we do, it’s always with the intention of modernizing and adapting to more cloud-native architectures. Consequently, our clients are not usually looking to merely replicate their current org structure in the cloud. Instead, they’re looking to tailor it appropriately.

Defining Lines of Responsibility

What should developers need to understand and be responsible for? There tend to be two schools of thought at two different extremes when it comes to this depending on peoples’ backgrounds and experiences. Oftentimes, developers will want more control over infrastructure and operations, having come from the constraints of a more siloed organization. On the flip side, operations folks and managers will likely be more in favor of having a separate group retain control over production environments and infrastructure for various reasons—efficiency, stability, and security to name a few. Not to mention, there are a lot of operational concerns that many developers are likely not even aware of—the sort of unsung, unglamorous bits of running software.

Ironically, both models can be used as an argument for “DevOps.” There are also cases to be made for either. The developer argument is better delivery velocity and innovation at a team level. The operations argument is better stability, risk management, and cost control. There’s also likely more potential for better consistency and throughput at an organization level.

The answer, unsurprisingly, is a combination of both.

There is an inherent tension between empowering developers and running an efficient organization. We want to give developers the flexibility and autonomy they need to develop good solutions and innovate. At the same time, we also need to realize the operational efficiencies that common solutions and standardization provide in order to benefit from economies of scale. Should every developer be a generalist or should there be specialists?

Real Kinetic helps clients adopt a model we refer to as “Developer Enablement.” The idea of Developer Enablement is shifting the focus of ops teams from being “masters” of production to “enablers” of production by applying a product lens to operations. In practical terms, this means less running production workloads on behalf of developers and more providing tools and products that allow developers to run workloads themselves. It also means thinking of operations less as a task-driven service model and more as a strategic enabler. However, Developer Enablement is not about giving full autonomy to developers to do as they please, it’s about providing the abstractions they need to be successful on the platform while realizing the operational efficiencies possible in a larger organization. This means providing common tooling, products, and patterns. These are developed in partnership with product teams so that they meet the needs of the organization. Some companies might refer to this as a “platform” team, though I think this has a slightly different meaning. So how does this map to an actual organization?

Mapping Out an Engineering Organization

First, let’s mentally model our engineering organization as two groups: Product Development and Infrastructure and Reliability. The first is charged with developing products for end users and customers. This is the stuff that makes the business money. The second is responsible for supporting the first. This is where the notion of “developer enablement” comes into play. And while this group isn’t necessarily doing work that is directly strategic to the business, it is work that is critical to providing efficiencies and keeping the lights on just the same. This would traditionally be referred to as Operations.

As mentioned above, the focus of this discussion is the green box. And as you might infer from the name, this group is itself composed of two subgroups. Infrastructure is about enabling product teams, and Reliability is about providing a first line of defense when it comes to triaging production incidents. This latter subgroup is, in and of itself, its own post and worthy of a separate discussion, so we’ll set that aside for another day. We are really focused on what a cloud infrastructure organization might look like. Let’s drill down on that piece of the green box.

An Infrastructure Organization Model

When thinking about organization structure, I find that it helps to consider layers of operational concern while mapping the ownership of those concerns. The below diagram is an example of this. Note that these do not necessarily map to specific team boundaries. Some areas may have overlap, and responsibilities may also shift over time. This is mostly an exercise to identify key organizational needs and concerns.

We like to model the infrastructure organization as three teams: Developer Productivity, Infrastructure Engineering, and Cloud Engineering. Each team has its own charter and mission, but they are all in support of the overarching objective of enabling product development efficiently and at scale. In some cases, these teams consist of just a handful of engineers, and in other cases, they consist of dozens or hundreds of engineers depending on the size of the organization and its needs. These team sizes also change as the priorities and needs of the company evolve over time.

Developer Productivity

Developer Productivity is tasked with getting ideas from an engineer’s brain to a deployable artifact as efficiently as possible. This involves building or providing solutions for things like CI/CD, artifact repositories, documentation portals, developer onboarding, and general developer tooling. This team is primarily an engineering spend multiplier. Often a small Developer Productivity team can create a great deal of leverage by providing these different tools and products to the organization. Their core mandate is reducing friction in the delivery process.

Infrastructure Engineering

The Infrastructure Engineering team is responsible for making the process of getting a deployable artifact to production and managing it as painless as possible for product teams. Often this looks like providing an “opinionated platform” on top of the cloud provider. Completely opening up a platform such as AWS for developers to freely use can be problematic for larger organizations because of cost and time inefficiencies. It also makes security and compliance teams’ jobs much more difficult. Therefore, this group must walk the fine line between providing developers with enough flexibility to be productive and move fast while ensuring aggregate efficiencies to maintain organization-wide throughput as well as manage costs and risk. This can look like providing a Kubernetes cluster as a service with opinions around components like load balancing, logging, monitoring, deployments, and intra-service communication patterns. Infrastructure Engineering should also provide tooling for teams to manage production services in a way that meets the organization’s regulatory requirements.

The question of ownership is important. In some organizations, the Infrastructure Engineering team may own and operate infrastructure services, such as common compute clusters, databases, or message queues. In others, they might simply provide opinionated guard rails around these things. Most commonly, it is a combination of both. Without this, it’s easy to end up with every team running their own unique messaging system, database, cache, or other piece of infrastructure. You’ll have lots of architecture astronauts on your hands, and they will need to be able to answer questions around things like high availability and disaster recovery. This leads to significant inefficiencies and operational issues. Even if there isn’t shared infrastructure, it’s valuable to have an opinionated set of technologies to consolidate institutional knowledge, tooling, patterns, and practices. This doesn’t have to act as a hard-and-fast rule, but it means teams should be able to make a good case for operating outside of the guard rails provided.

This model is different from traditional operations in that it takes a product-mindset approach to providing solutions to internal customers. This means it’s important that the group is able to understand and empathize with the product teams they serve in order to identify areas for improvement. It also means productizing and automating traditional operations tasks while encouraging good patterns and practices. This is a radical departure from the way in which most operations teams normally operate. It’s closer to how a product development team should work.

This group should also own standards around things like logging and instrumentation. These standards allow the team to develop tools and services that deal with this data across the entire organization. I’ve talked about this notion with the Observability Pipeline.

Cloud Engineering

Cloud Engineering might be closest to what most would consider a conventional operations team. In fact, we used to refer to this group as Cloud Operations but have since moved away from that vernacular due to the connotation the word “operations” carries. This group is responsible for handling common low-level concerns, underlying subsystems management, and realizing efficiencies at an aggregate level. Let’s break down what that means in practice by looking at some examples. We’ll continue using AWS to demonstrate, but the same applies across any cloud provider.

One of the low-level concerns this group is responsible for is AMI and base container image maintenance. This might be the AMIs used for Kubernetes nodes and the base images used by application pods running in the cluster. These are critical components as they directly relate to the organization’s security and compliance posture. They are also pieces most developers in a large organization are not well-equipped to—or interested in—dealing with. Patch management is a fundamental concern that often takes a back seat to feature development. Other examples of this include network configuration, certificate management, logging agents, intrusion detection, and SIEM. These are all important aspects of keeping the lights on and the company’s name out of the news headlines. Having a group that specializes in these shared operational concerns is vital.

In terms of realizing efficiencies, this mostly consists of managing AWS accounts, organization policies (another important security facet), and billing. This group owns cloud spend across the organization and, as a result, is able to monitor cumulative usage and identify areas for optimization. This might look like implementing resource-tagging policies, managing Reserved Instances, or negotiating with AWS on committed spend agreements. Spend is one of the reasons large companies standardize on a single cloud platform, so it’s essential to have good visibility and ownership over this. Note that this team is not responsible for the spend itself, rather they are responsible for visibility into the spend and cost allocations to hold teams accountable.

The unfortunate reality is that if the Cloud Engineering team does their job well, no one really thinks about them. That’s just the nature of this kind of work, but it has a massive impact on the company’s bottom line.

Summary

Depending on the company culture, words like “standards” and “opinionated” might be considered taboo. These can be especially unsettling for developers who have worked in rigid or siloed environments. However, it doesn’t have to be all or nothing. These opinions are more meant to serve as a beaten path which makes it easier and faster for teams to deliver products and focus on business value. In fact, opinionation will accelerate cloud adoption for many organizations, enable creativity on the value rather than solution architecture, and improve efficiency and consistency at a number of levels like skills, knowledge, operations, and security. The key is in understanding how to balance this with flexibility so as to not overly constrain developers.

We like taking a product approach to operations because it moves away from the “ticket-driven” and gatekeeper model that plagues so many organizations. By thinking like a product team, infrastructure and operations groups are better able to serve developers. They are also better able to scale—something that is consistently difficult for more interrupt-driven ops teams who so often find themselves becoming the bottleneck.

Notice that I’ve entirely sidestepped terms like “DevOps” and “SRE” in this discussion. That is intentional as these concepts frequently serve as a distraction for companies who are just beginning their journey to the cloud. There are ideas encapsulated by these philosophies which provide important direction and practices, but it’s imperative to not get too caught up in the dogma. Otherwise, it’s easy to spin your wheels and chase things that, at least early on, are not particularly meaningful. It’s more impactful to focus on fundamentals and finding some success early on versus trying to approach things as town planners.

Moreover, for many companies, the organization model I walked through above was the result of evolving and adapting as needs changed and less of a wholesale reorg. In the spirit of product mindset, we encourage starting small and iterating as opposed to boiling the ocean. The model above can hopefully act as a framework to help you identify needs and areas of ownership within your own organization. Keep in mind that these areas of responsibility might shift over time as capabilities are implemented and added.

Lastly, do not mistake this framework as something that might preclude exploration, learning, and innovation on the part of development teams. Again, opinionation and standards are not binding but rather act as a path of least resistance to facilitate efficiency. It’s important teams have a safe playground for exploratory work. Ideally, new ideas and discoveries that are shown to add value can be standardized over time and become part of that beaten path. This way we can make them more repeatable and scale their benefits rather than keeping them as one-off solutions.

How has your organization approached cloud development? What’s worked? What hasn’t? I’d love to hear from you.