More than 90 percent of enterprises are using or plan to use microservices and 73 percent said they found it harder to report problems with applications performance in a microservices environment. These findings are part of a microservices report written by Dimensional Research and commissioned by monitoring company LightStep.
The report defined a microservice as “a small, independently deployable, independently scalable software service that is designed to encapsulate a specific semantic function in the larger application.” Many enterprises are adopting these to achieve both agility and scalability, according to the findings. Smaller drivers were performance improvements and to increase accountability within the organization.
For the report, 353 senior development stakeholders, who are involved in microservices adoption in company with more than 500 employees, participated in an online survey. However, participants were asked first if they had plans for microservices and if they answered no (which 32 did), they were excluded from the remainder of the report.
While the growth of microservices is a relatively new phenomenon, the momentum of this architecture doesn’t appear to be going anywhere. 98 percent of those surveyed expect for it to become the default architecture and 86 percent believe it will happen in the next five years.
Pete Abrams, the co-founder and COO of microservices application performance monitoring (APM) startup Instana, agrees that this will be the “new default” and attributes it to the digital transformation in the marketplace and the need for organizations to build new software, and build it faster.
Microservices are well equipped to enable this shift to building new software, but it has challenges. “The hardest step is the agile step. You have to reprogram your humans on how to think about their process for constructing software. Once you do that, then the microservice cloud container orchestration sequence is pretty fluid; it’s pretty natural,” said Abrams.
The report reflects this notion. It found that 99 percent of those surveyed faced challenges with their microservices architecture. The problems included an increase in operational burdens; inexperience and lack of skills from teams to build, leverage, and manage microservices; and increased difficulty in finding and resolving performance issues, which has impact on business.
Microservices architectures are more difficult to monitor because there is an increase in application data, and traditional monitoring tools are not designed for the complexity of microservices.
However, these challenges aren’t creating a barrier to adoption. “Microservices enable organizations to be more agile and scale more effectively,” said Ben Sigelman, LightStep co-founder and CEO. “Monolithic applications can create a logjam for developers because teams cannot operate independently, making it extremely difficult to innovate and scale effectively.”
So it falls to APM companies, like LightStep and Instana, to create tools to troubleshoot problems in the new microservices architecture.
LightStep’s own APM tool, [x]PM, was designed for microservices and traces and monitors across service boundaries without upfront data sampling, agents, or performance overhead.
To address the influx of application data, LightStep [x]PM deploys a decentralized architecture that can analyze every transaction in each service in production and measure performance.
Another key element of the APM tool is that it can monitor beyond applications into all layers of the infrastructure. This means end-users can monitor “what matters to them most,” according to Sigelman, as well as the complete distributed application “from mobile to web clients, to microservices and monoliths.”
Additionally, LightStep has an agentless design and relies on open standards for instrumentation, can access API data, and allows users to identify service level objectives.
LightStep was founded by three ex-Googlers. The company launched [x]PM and announced its Series A funding in November 2017.
Instana’s APM tool also was designed specifically for microservices architectures, including its applications and containers. Today, it extended its monitoring capabilities to include Kubernetes.
The tool has three core capabilities. First, by building an internal data model it offers continuous discovery and visualization of the entire technology stack to identify performance issues. Second, the APM tool monitors and visualizes this data in real-time. And lastly, it offers artificial intelligence (AI) capabilities, which Abrams said is a necessity due to the complexity of these environments.
Abrams calls this need for APM-type information across all organization and infrastructure layers, “democratization.”
Its newly-added capabilities for Kubernetes monitors its health as well as the orchestrated applications and systems and the interactions and interdependencies between the infrastructure and the Kubernetes software. It discovers, monitors, and analyzes Kubernetes labels, clusters, pods, and the orchestrated application.
The company also announced this week that the U.S. subsidiary of Yahoo Japan Actapio chose Instana to monitor its hybrid environment, which includes Kubernetes clusters and infrastructure technology such as OpenStack.
Instana is based out of Silicon Valley and was founded by Mirko Novakovic, CEO, Pavlo Baron, CTO, Fabian Lange, VP of engineering, and Abrams — all of whom have been in APM for decades, including Abrams who was the seventh employee at AppDynamics, which is now a part of Cisco.