Published: August 6, 2020
The 451 Take
It's time for a change
Each of the components of this dynamic environment generates operations data, as does the communications between the services that make up these distributed applications, resulting in a much larger volume of IT operations data than that generated by traditional applications. To learn that a performance problem is occuring and quickly and accurately identify the cause of a problem, teams must collect and analyze this potentially enormous volume of data. This new reality requires a new approach to monitoring that in many cases includes new monitoring tools.
However, new tools should not necessarily mean additional tools. We commonly hear from organizations that face a number of important challenges because they are juggling too many tools that collect and analyze only a subset of relevant operations data.
Our recent survey indicates that users either already have or plan to implement a very wide array of monitoring functionalities. We asked respondents about their use of 12 different types of commonly used monitoring and incident response tools, and found strong current and planned adoption of all 12. The percentage of respondents who had no plans to deploy the tools was remarkably small, with network monitoring and database monitoring each registering zero respondents with no plans to deploy, and log management and infrastructure monitoring with only 1% each with no plans to deploy.
These results do not indicate that respondents necessarily plan to use a different tool from a different vendor for each function, and we would recommend that organizations consider using tools that offer multiple functions. It's increasingly possible to do so, with many of the larger vendors expanding horizontally over the past few years to deliver several types of tools. In fact, among the top 10 vendors by revenue in this sector, all have offerings in six or more of the categories we track, with two delivering in eight categories, according to our research.
Integrated tools can solve a few challenges for organizations. One is that they simplify vendor and tool management. More importantly, integrated tools have the potential to deliver new capabilities that can address some of the problems that commonly plague organizations that are adopting cloud-native technologies. One of the most common integrations we've seen over the past few years is infrastructure monitoring and log analytics, where a user discovers that a problem is occurring via the metrics-centric infrastructure monitoring functionality but can view within the same visualization relevant logs that may help pinpoint the root cause of the problem.
Those capabilities are increasingly combined with distributed tracing data, which can help narrow down trouble spots and contribute to detailed topology maps. Modern tools can run sophisticated analytics across metrics, distributed traces, logs and events to group together related issues and guide users to the source of a problem and how to solve it. Such functionality can have a significant impact on incident response by reducing or eliminating alert storms and decreasing the time it takes to identify and fix performance issues.
Nancy Gohring is a senior research analyst for the Applied Infrastructure & DevOps and Cloud Native Channels at 451 Research, a part of S&P Global Market Intelligence. Nancy follows the monitoring and incident response markets including APM, infrastructure monitoring, log analytics, distributed tracing and event analytics. Her coverage encompasses recent trends around observability and AIOps, both of which are influenced by enterprise adoption of cloud-native technologies.