The State of Observability in 2021By George Miranda | Last modified on August 27, 2021
Observability adoption has increased as more companies seek to understand how their applications behave in production and quickly identify and resolve problems. Our second annual observability maturity report is the first that shows a year-over-year overview of that growth in adoption.
Our observability report aims to understand community perceptions and awareness of observability, how engineering teams approach observability, and then map an observability maturity model that reflects current research findings.
The observability maturity model focuses on team capabilities and outcomes. It goes beyond what types of data teams collect (e.g., logs, metrics, and traces) and focuses on what kinds of problems they’re able to solve.
The observability maturity spectrum
Our observability maturity report shows that while more people are adopting observability, they seem to be at the early stages of the maturity spectrum. Our observability maturity spectrum recognizes five unique groups:
1. No Plans
The organizations that fall under this group have no plans to adopt observability in their systems. Less than one-fifth (17%) of our respondents fall into this category, and they help form the foundation of our expectations when compared with other groups. More than half of this 17% are financial companies with more than 1,000 employees.
This group has indicated some interest in observability and has plans to adopt the practice in a year or two. About 26% of the respondents in this group are beginning to practice observability on a team-by-team basis. At the same time, about 33% intend to do so within the following year. Overall, about 20% of teams surveyed do not have tooling in place to practice observability, but they have plans to do so in the next 12 months.
The Novice group comprises teams at the earliest stage of observability. They either report practicing some observability processes or have some tooling system for observability. However, in rare cases, they report having both. They also report some key capabilities, such as identifying and resolving bugs before and after deploying to production. This group houses most of our respondents, with 37% falling under this category.
About 13% of our respondents are in this group. They’re starting to grasp how their applications operate and behave in production. These teams report procedures, tooling, and results consistent with relatively sophisticated observability practices. This group also saw a 4% decrease from last year, with that small collective moving on to the Advanced group.
The Advanced group understands their entire system and is reaping the benefits of practicing observability. They have some observability tooling systems and follow practices that ensure their system is highly observable. Only 14% of our respondents are in this category.
The top tiers of this spectrum (Intermediate and Advanced) have significantly increased their confidence in resolving errors in the past year. Of these two groups, half have adopted observability across their entire organization, while 43% are adopting on a team-by-team basis.
The overall proportion of these top categories remained the same, signaling that lower maturity groups (No Plans and Planning) are stagnating in their journey. That could suggest that teams who are well versed in observability practices are accelerating their skills and may be pulling away from their lower-maturity counterparts.
The benefits of practicing observability
Based on our 2021 observability report data, teams in the upper tiers of our maturity spectrum realize the following benefits:
1. Higher productivity
Our observability report indicates that 69% of the teams who have adopted observability can immediately identify when a problem arises and its impact on other systems. Even better, more than half (51%) of those teams can immediately identify a solution to the problem.
2. Improvement in code quality
Observability helps the majority of teams (63%) understand their entire application system, including events and high-level trends. These teams are then able to write and maintain high-quality code.
3. Higher end-user satisfaction
Virtually all respondents in the Advanced group and almost 90% in the Intermediate group indicated high customer satisfaction rates. They were also 3X more likely to have customers who were always satisfied than groups who had not yet started practicing observability.
4. Software developer retention
Most (81%) of the Intermediate and Avanced teams report confidence in catching bugs before production, while 77% of No Plans teams report little to no confidence in their ability to catch bugs before production. With the confidence level of the Intermediate and Advanced groups, software developers can ship code faster into production. As a result, they spend more time building exciting new features and are less likely to leave a company due to frustrations with the application system.
Our observability report also confirms that those in the higher tiers of the maturity spectrum may have:
- The ability to follow predictable repeated cycles
- Mostly automated or completely automated releases
- The ability to set and measure service-level objectives
- The ability to picture context-rich events
Other key findings in our observability report
In our research on the adoption of observability in organizations, we also found:
- Software developers are the primary drivers of observability, followed closely by DevOps engineers and site reliability engineers (SREs). Some respondents also identified themselves as the drivers of observability in their organization, with 21% of them being developers, 36% DevOps or SRE, 14% managers, 16% architects, 10% CIO/CTO, and 3% operations.
- Lower-maturity teams already practicing observability but not realizing the benefits also disproportionately cite a lack of implementation skills as one of their most significant barriers to adoption. When this barrier combines with other factors—like no tooling system—we believe this signals either confusion in the market over which capabilities should be achievable when practicing observability or that teams need more training to achieve those capabilities (or both).
- About 8% have moved from the planning phase into actual practice since last year. That increase occurs mainly across individual teams (7%) instead of the entire organization (1%).
About this research
We conducted the survey used to generate these findings from December 2020 – January 2021 with ClearPath Strategies, a boutique consulting firm recommended to Honeycomb by Dr. Nicole Forsgren, former lead researcher for DORA and the State of DevOps reports. We distributed invites to participate in the survey via a Honeycomb email list and social media outreach, and a total of 405 respondents participated.
Download and read the full observability report to know how mature teams get the results they do, what those just getting started can do to avoid stagnating in the middle, and how to tell if your observability adoption initiative is heading in the right direction.
Have questions about how you can improve your system’s observability? Sign up for office hours with one of our developer advocates!
Ask Miss O11y: Is There a Beginner’s Guide On How to Add Observability to Your Applications?
Dear Miss O11y, I want to make my microservices more observable. Currently, I only have logs. I’ll add metrics soon, but I’m not really sure...
Feature Focus: Winter Edition ❄️
It’s been a minute since our last Feature Focus, and we have a bit of catching up to do! I’m happy to report we’ll resume...
Iterating Our Way Toward a Service Map
For a long time at Honeycomb, we envisioned using the tracing data you send us to generate a service map. If you’re unfamiliar, a service...