Splunk, an arm of Cisco, has donated an OpenTelemetry Injector library it developed for streamlining instrumentation of applications that have not been containerized, to the open source project that is being advanced under the auspices of the Cloud Native Computing Foundation (CNCF).
Greg Leffler, director of developer evangelism for Splunk, said this latest contribution to the project will make it simpler for DevOps teams with zero touch to instrument legacy applications that have already been deployed in a production environment. OpenTelemetry Injector achieves that goal by bundling a binary with the OpenTelemetry auto-instrumentation agents for specific runtime environments.
Written in the Zig programming language, that capability enabled by OpenTelemetry Injector is crucial because most organizations continue to run legacy monolith applications that have yet to be encapsulated into a set of containers, he added.
A survey of 1,855 IT operations and engineering professionals suggests the business benefits enabled by instrumentation applications are well worth the effort. The survey finds organizations that have adopted OpenTelemetry are seeing positive effects on their revenue growth (72%), operating margins (71%), and brand perception (71%) specifically because of it.
Nearly three quarters (74%) of observability practitioners indicate observability has a positive impact on employee productivity, while 69% report it has a positive impact on customer experience. Nearly two thirds also said it has had a positive impact on overall revenue, with 74% noting that being able to monitor critical business processes is at least moderately important to their organization.
Nearly half (47%) said alerts generated by observability platforms significantly influence security decisions within their organization. Additionally, 64% said they encounter fewer application and infrastructure performance issues thanks to better collaboration with security teams, with 62% reporting that they troubleshoot and solve issues with their security team. However, 59% also identified resistance to change as the biggest barrier to improving collaboration.
Going forward, organizations will also need more telemetry data to successfully operationalize artificial intelligence (AI), noted Leffler. In fact, the survey finds nearly half of respondents (48%) identify low data quality as the main barrier to AI readiness, with 47% noting that monitoring AI workloads has made their jobs more challenging.
On the plus side, more than three quarters (78%) said AI has enabled them to spend more time on innovation than maintenance.
It’s not clear when OpenTelemetry will be extended to support schemas needed to observe AI applications. Splunk in collaboration with AGNTCY, a collective dedicated to the discovery, identity, messaging, and observability for AI agents, donated a schema to OpenTelemetry that allows users to emit data about LLMs in a consistent way. The goal is to make it possible to analyze large language model (LLM) outputs on multiple levels using qualitative metrics to validate right answers and identify bias and quantitative metrics that track cost and performance.
It’s not clear how pervasively OpenTelemetry instrumentation tools have been deployed, but it’s clear that it is increasingly being used beyond DevOps workflows to collect a range of telemetry data. The challenge now is turning all that data into actual actionable intelligence.
