Cloud-native applications built using technologies such as containers and serverless computing framework are now the primary drivers of application innovation. A global survey of 1,140 IT professionals working for organizations with more than $500 million in revenue finds that nearly half (49%) of their innovation initiatives are being delivered with cloud native technologies, with that percentage expected to increase to 58% in the next five years.
Conducted by Cisco AppDynamics, the survey also notes 83% of respondents are also finding that cloud native technologies lead to increased complexity within IT environments that are becoming more hybrid by the day as cloud native applications built employing microservices with lots of dependencies are deployed alongside legacy monolithic applications.
Other significant challenges that survey respondents identified as IT environments become more hybrid include expanding attack surfaces (42%) visibility gaps into application performance (41%), aligning cloud costs to performance (39%) and increased complexity caused by microservices and containers (36%).
That issue is also contributing to significant amounts of churn with organizations. The survey finds well over a third (36%) of respondents reporting that silos and ineffective collaboration are already resulting in IT talent leaving their organization, with 46% predicting that churn will increase if silos continue to persist.
As a result, 85% of respondents now view observability as a strategic priority for their organization, with more than half reporting that their organization is already exploring their options, with more than three-quarters (78%) are now collecting metrics, events, logs and traces at rates across hybrid cloud computing environments that makes legacy approaches to monitoring IT environments based solely on pre-defined metrics impractical.
On the plus side, cloud-native applications generate a lot more telemetry data than traditional monolithic applications so armed with the right observability tools it’s possible to trouble shoot these applications by launching queries that surface the root cause of any issue. The immediate issue that IT teams are encountering is the level of expertise required to craft those queries are, as always, in short supply.
It’s still early days as far as the adoption of observability platforms is concerned, but theoretically machine learning algorithms will hopefully soon be widely used to reduce the level of DevOps expertise that would otherwise be needed to craft those queries. Instead, machine learning algorithms will leverage artificial intelligence (AI) models to predict when those issues are likely to impact application performance.
The issue, of course, is that for the moment at least the deployment of cloud-native applications is outpacing the rate at which many DevOps teams can fund the acquisition of observability platforms infused with machine learning algorithms that promise to augment DevOps teams that are as more microservices become distributed across the enterprise are experiencing higher levels of strain.
It’s not clear to what degree in addition to deploying cloud-native applications that IT organizations will also move to use the same technologies to modernize legacy applications. Many IT teams, for example, are finding it more efficient to isolate services that are embedded with a larger application into a microservice that multiple applications can invoke by calling an application programming interface (API). The one thing that is certain is microservices will soon be pervasively distributed across the enterprise in a way that will require more advanced tooling to successfully manage.
In the meantime, IT leaders would be well-advised to assess the pace at which cloud-native applications that are under development today are likely to be deployed in production environments. That forecast should give them a good idea of how much time they might have to acquire tools such as observability platforms before reaching a level of cloud-native application deployment that given the current number of applications already running is unsustainable. After all, no matter how great any application is the one thing it won’t do any time soon is manage itself.