Open SDK Leveraged as Industry Shifts Toward Agent-Driven Operations
The News
At KubeCon North America 2025, Last9 discussed rapid advancement in agentic automation for observability and remediation, highlighting marked growth in adoption and trust over the past six months as organizations move from skepticism to confidence in agent-based solutions. The company released an agent tech SDK enabling organizations to automate remedial tasks and close the loop from observability to automated test case generation and code fixes, emphasizing that data quality serves as the foundation for successful agentic automation.
Past challenges with schema differences and data quality in other tools hindered adoption, but addressing these issues enables production deployment with predictable and reliable outcomes. Last9 announced plans to open-source their SDK to drive broader adoption and ecosystem participation through community-driven development, positioning open agent architectures as the next frontier analogous to open-source Grafana’s rise for dashboards in the 2010s. The company emphasizes interoperability and open protocols like OpenTelemetry as critical for data ingestion and increasingly for consumption and remediation layers, predicting that open agent architectures will become the norm for automation and remediation to ensure cross-platform interoperability.
Analyst Take
Last9 projects that agentic operations will become mainstream by 2026, transcending traditional AIOps and MLOps that often remain deterministic, with agentic approaches offering non-deterministic, hypothesis-driven capabilities enabling more dynamic and effective remediation and operational management. The company describes an acceleration from waterfall to agile to DevOps toward continuous agent-driven automation, with feedback loops compressing from weeks to hours or minutes in a paradigm of “extreme agentic automation” where agents detect, remediate, and optimize systems in real time. Despite increasing automation capabilities, Last9 emphasizes that human accountability remains essential, citing incidents like AI-driven errors in customer-facing systems that underscore the need for checks, balances, and clear responsibility regardless of automation level. The company frames automation as analogous to power tools versus manual tools; productivity rises, but responsibility remains with the user, with LLMs and agents enabling adaptive automation that surpasses static scripts in dynamic modern systems while still requiring oversight to ensure quality and safety.
Last9’s emphasis on data quality as the foundation for agentic automation addresses a critical but often overlooked prerequisite: agents can only be as reliable as the data they consume, and observability data quality remains inconsistent across many organizations.
Schema differences, incomplete instrumentation, inconsistent tagging, and telemetry gaps create scenarios where agents make decisions based on partial or inaccurate information, leading to incorrect remediation actions that potentially cause greater harm than the original issues. The company’s focus on addressing data quality before deploying agents reflects operational maturity, but it also reveals a chicken-and-egg challenge: organizations need high-quality observability data to benefit from agents, but achieving that data quality often requires the automation and consistency that agents promise to deliver. Organizations must determine whether to invest in data quality improvement as a prerequisite to agent adoption or adopt agents incrementally in well-instrumented environments while gradually expanding coverage.
The open-sourcing of Last9’s agent SDK and positioning of open agent architectures as analogous to Grafana’s open-source dashboard success reflects strategic calculation about ecosystem dynamics. Grafana succeeded by becoming the de facto standard through community adoption and extensibility, creating network effects where visualization plugins, integrations, and expertise coalesced around a single platform.
Last9 appears to pursue a similar positioning for agent-based automation, betting that open standards and community participation will drive adoption more effectively than proprietary solutions. However, the analogy has limits—dashboards are primarily consumption interfaces where standardization benefits users through consistency and portability, while agents perform actions with operational consequences where standardization may constrain flexibility and differentiation. The market will determine whether open agent architectures create the same network effects as open-source dashboards or whether the operational risk and customization requirements of automation favor proprietary, tightly integrated solutions.
The prediction that agentic operations will transcend traditional AIOps and MLOps by 2026 reflects optimism about agent capabilities that current production deployments have not yet validated at scale. Traditional AIOps uses machine learning for anomaly detection, correlation, and pattern recognition, but typically requires human decision-making for remediation actions.
Agentic approaches promise to close this loop with autonomous remediation, but the non-deterministic, hypothesis-driven nature that Last9 emphasizes also introduces unpredictability that operations teams historically avoid in production environments. Our Day 2 research found that 84.5% of organizations have adopted or are adopting AI-powered issue detection, and 80.5% are pursuing performance optimization, but these capabilities focus on insights rather than autonomous actions.
The gap between AI-powered detection and agent-driven remediation represents a significant trust and reliability threshold that organizations must cross, requiring evidence that agents consistently make correct decisions across diverse failure scenarios rather than occasionally producing catastrophic errors.
The emphasis on human accountability despite increasing automation reflects necessary realism about responsibility and liability when automated systems fail. The comparison to power tools, where productivity increases but responsibility remains with the user, provides useful framing, but it also understates the autonomy gap between tools that amplify human actions and agents that take independent actions. When a power tool causes damage, the operator’s decisions directly lead to the outcome; when an agent causes an outage, the relationship between human oversight and agent autonomy becomes less clear.
Organizations deploying agentic automation must establish clear accountability frameworks that define when humans should intervene, what approval workflows govern different action types, and how to audit agent decisions when incidents occur. The feedback loop compression from weeks to minutes that Last9 describes as “extreme agentic automation” creates scenarios where human oversight becomes impractical, forcing organizations to either accept agent autonomy with associated risks or constrain automation to preserve human control.
Looking Ahead
Last9’s success with agentic automation depends on whether the open SDK and agent architecture strategy attracts sufficient ecosystem participation to create network effects and establish industry standards. Open-sourcing the SDK reduces Last9’s direct control over the technology’s evolution but potentially accelerates adoption if community contributions add capabilities and integrations that a single vendor could not develop alone.
The next 12-18 months will reveal whether observability and automation vendors, cloud providers, and enterprises contribute to Last9’s open agent architecture or pursue competing approaches, either proprietary solutions that maintain differentiation or alternative open standards that fragment the ecosystem. The company’s challenge is building enough initial momentum and demonstrating enough value that the open architecture becomes the default choice rather than one option among many.
The broader shift toward agentic operations that Last9 projects represents a fundamental change in how organizations approach infrastructure management, but the timeline and adoption curve remain highly uncertain. The compression of feedback loops from weeks to minutes creates operational efficiency but also introduces brittleness; systems optimized for rapid automated response may lack resilience when automation fails or makes incorrect decisions.
Organizations must balance the productivity gains of extreme automation against the operational risk of reduced human oversight and the potential for cascading failures when agents make errors at machine speed. The industry’s experience with previous automation waves, from configuration management to infrastructure as code to GitOps, suggests that adoption follows a pattern where early adopters experiment aggressively, high-profile failures create caution, and eventual mainstream adoption occurs with guardrails and governance that limit both benefits and risks. Agentic automation will likely follow similar patterns, with the outcome depending on whether the technology matures sufficiently to deliver consistent reliability before skepticism from early failures limits broader adoption.
High Availability Becomes a Strategic Control Layer in 2026 IT Architectures
NKP 2.17 Signals a Security-First Turn in Enterprise Kubernetes Platforms
Omni Agent Intelligence Signals the Next Phase of AI-Native CX Governance
Always-On Retail Becomes a Platform Requirement at NRF 2026
Agentic Commerce Moves From Concept to Control Plane at NRF 2026
