The Fallacy of “Full-Stack” Observability Tools

The term "full-stack" observability has been thrown around for quite some time, promising fewer outages, five-nines uptime, and reduced MTTR. Many observability tools and platforms have been quick to capitalize on this trend and bill themselves as “full-stack observability tools” that can deliver a holistic view of the health of your system and all its components.

While that might be true for web apps, this term is a misnomer for mobile apps. This post will explore why most “full-stack observability tools” just don’t cut it for mobile teams.

‍

‍

Full-stack observability platforms will be quick to point out their support for mobile apps but more often than not, the level of that support leaves much to be desired. They might be reporting some telemetry from the mobile app, but in most cases, much more is needed to make your mobile app observable.

‍

Web vs. mobile: apples and oranges

It’s not just a matter of the depth and breadth of the telemetry. Mobile development is a paradigm shift from web development and needs a different, better-tailored approach.

Full-stack observability typically refers to monitoring web applications, encompassing the front-end, back-end, and infrastructure. This works well when you’re running your apps on your servers and delivering them through your network and infrastructure.

On the other hand, mobile apps run several different versions on a huge variety of environments and devices to which you do not have access. Mobile teams are not purely front-end and manage complex state logic, data persistence, and other "backend-like" functionalities – all while remaining distinct from centralized backend teams. Often, these backend services belong to separate "full-stack" teams responsible for web development, leaving mobile developers in a silo.

‍

Full-stack for web, not mobile

The truth is that most vendors who bill themselves as full-stack observability providers focus on the world of web development – back-end services, infrastructure, and front-end experiences. Mobile, with its unique challenges, gets left behind.

This massive disconnect between full-stack observability and mobile development realities is sadly much too common. In fact, in a conversation with an Ops leader at a top software company responsible for "full-stack observability", I was shocked when I learned they have never interacted with a mobile developer in their 13 years with the company.

‍

What mobile teams do need

So, what do mobile teams need to achieve true observability of their apps? Here are some key aspects:

1. Release perspective: as mentioned earlier, mobile apps always have several versions running at any given time. Aggregating telemetry by time instead of the app version can mask emerging issues in your newest and most-adopted versions. Mobile teams need tools that adopt this perspective and tailor their approach to mobile development.
2. User-centric mobile health metrics: The legacy health and performance metrics used for web development can’t accurately represent the vastly different mobile user experience. Mobile teams need centralized dashboards with mobile-specific health metrics (crashes, ANRs, app launch, loading time, etc.) to enable the developers to identify areas across the entire stack requiring attention.
3. The right tools and seamless handoff: While there is merit to using a single platform for your entire stack, you first need to ensure it can actually deliver observability over every part of your stack. This is not currently an option for mobile teams.
   Mobile teams need tools that facilitate a smooth debugging experience by starting with user experience issues and seamlessly transitioning to the relevant backend service data when needed.

‍

‍

All-in-one full-stack observability solutions are not currently sufficient to satisfy the unique requirements of mobile, and won’t be any time soon. Mobile teams have to choose between achieving full-stack observability with more than one tool or sacrificing mobile observability for an all-in-one solution.

Successful mobile teams understand that this is not an acceptable trade-off and demand tools that cater to the specific needs of mobile development. This provides a holistic view of their mobile app's health, from user experience to backend interactions.