July 2026 Summaries

Grafana Labs has developed Agent Observability, a tool now available on Grafana Cloud, to address the challenges of monitoring and scaling agent workloads, especially those involving advanced AI models. Initially created to manage the complexities of Grafana Assistant's rapid growth, the tool provides a structured approach for observing agent behavior, triaging issues, and improving performance through phases of development. It utilizes traditional engineering metrics alongside innovative evaluation techniques like LLM-judge evaluators and deterministic checks to ensure agent reliability. The tool supports offline evaluation, CI/CD integration, and ongoing experimentation, enabling developers to refine their agents with confidence. Agent Observability enhances traditional software monitoring by offering a comprehensive framework for iterating on non-deterministic agents, helping developers respond effectively to the dynamic nature of AI-driven tasks.

AI Week at Grafana Labs highlighted the evolving intersection of AI and observability, emphasizing how these technologies are transforming problem-solving in software development and management. The event showcased various new AI-driven tools and features, such as Workspace, Assistant Investigations, and Automations, that enhance operational efficiency and agent observability. With AI's growing role in automating routine tasks and enabling more meaningful work, Grafana Labs is committed to innovation and user feedback to ensure that their tools address real-world challenges effectively. The company recognizes the rapid pace of change and aims to support users in adapting to these advancements, fostering a future where AI-driven insights lead to more informed decision-making. The event encouraged participants to engage with the tools and provide feedback, highlighting Grafana Labs' dedication to refining their offerings based on user experiences and needs.

Grafana Cloud's AI capabilities offer a transformative approach to managing the operational burdens of continuous integration and continuous delivery (CI/CD) by automating routine monitoring tasks and enhancing decision-making processes. By implementing AI-driven tools like Assistant Automations and Watchers, teams can streamline repetitive checks and continuously evaluate system health, allowing engineers to focus on applying their judgment where it is most valuable. These tools enable the identification and investigation of anomalies, ensuring that the operational loop not only automates routine tasks but also enhances the visibility and reliability of production systems. The integration of AI allows for a proactive approach to identifying issues that require attention, thereby reducing interruptions and enabling more informed release decisions. This approach aligns with modern expectations for AI transparency and human oversight, ensuring that while AI offloads repetitive vigilance tasks, accountability and critical decision-making remain with human practitioners.

Telemetry-driven development with AI tools like gcx and Grafana MCP is emphasized as a means to enhance confidence in code changes, especially when using coding agents. The narrative highlights the anxiety developers feel when merging AI-generated code due to a lack of familiarity, as opposed to manually written code. It suggests that leveraging telemetry data can bridge this gap by providing insights into how code behaves in real environments, thus aligning development speed with understanding. Tools such as gcx and Grafana MCP are presented as solutions that connect coding agents to telemetry data, enabling them to make informed decisions based on real-world metrics rather than assumptions. These tools allow for better observability and integration of system behaviors, improving the AI's ability to generate relevant solutions and maintain desired performance metrics. The text also touches on how AI can streamline tasks like traffic simulation and performance enhancements, offering a comprehensive feedback loop that ensures changes meet expected standards. The overall message advocates for a combination of AI and telemetry to maintain control and confidence in software development processes.

Grafana Assistant is an AI-driven tool integrated within Grafana Cloud, designed to make observability accessible from the early stages of software development by leveraging open-source technologies like Grafana, Prometheus, and OpenTelemetry. By enabling engineers to use natural language to interact with their system's telemetry data, Grafana Assistant simplifies the creation and management of dashboards and queries, reducing reliance on observability experts. The tool facilitates onboarding by providing contextual insights into existing systems and streamlines the planning of new services by suggesting relevant metrics and alerts. Integrating with broader software stacks, Grafana Assistant ensures observability is a foundational aspect of development, leading to quicker resolution of issues when they arise.

AI Week, hosted by Grafana, introduces a series of announcements and features aimed at enhancing observability in agentic operations, focusing on integrating AI tools into workflow management. With the rise of agents automating more processes and generating rapid changes, the shift emphasizes the need for observability to be incorporated from the outset, rather than as an afterthought. The event highlights new capabilities in Grafana Cloud, such as the Grafana Assistant, which is designed to aid in onboarding, planning, and incident management by leveraging telemetry data and AI-driven analysis to improve system understanding and performance. Participants can engage with the AI team through various platforms, explore technical demos, and use new tools like the Assistant Workspace and gcx command line tool to streamline operations. The initiative underscores the importance of maintaining system trust and collaboration amid the evolving landscape of agentic tooling, offering solutions like Agent Observability to monitor AI system behavior and enforce policies in real time. As part of the event, Grafana also introduces support for Microsoft Teams and an upgraded mobile app, ensuring that insights and collaboration are accessible across different communication channels.

Grafana Cloud has introduced a new allowlists design that consolidates its IP address management into a single, structured API, replacing the previous system of per-product lists and formats. This update simplifies the process of managing IP allowlists by providing a unified JSON schema that is regionally aware and easier to automate, offering better granularity and clarity in traffic direction. The legacy formats, including JSON lists, text equivalents, and DNS records, will remain operational until January 31, 2027, to facilitate a smooth transition. Users who rely on IP allowlisting must update their configurations to accommodate the new structured JSON format, while those using private connectivity like AWS PrivateLink may not need to make any changes. The change aims to enhance automation and clarity, allowing users to query specific regions or services without filtering through global lists, ensuring clearer application of allowlists across different deployment models.

Supabase has partnered with Grafana Labs to offer a seamless integration, allowing users to access Grafana Cloud's observability features directly from the Supabase dashboard with just one click. This integration enables developers to monitor key Supabase and Postgres metrics, enhancing visibility into query performance and database health, which is crucial as AI agents drive rapid application development. Supabase, an open-source managed Postgres platform, provides a comprehensive suite of backend services, making it easy for developers to deploy apps without sacrificing control. The collaboration with Grafana Cloud extends Supabase's open-source model to observability, offering a pre-configured dashboard that tracks essential metrics and identifies performance issues. This integration simplifies the observability strategy, allowing teams to focus on development while benefiting from production-grade insights and a free tier that supports most projects. The integration highlights Supabase's commitment to providing a great developer experience and Grafana Cloud's dedication to open standards, offering a scalable solution that grows with users' projects.

Grafana Cloud's cost attribution feature provides organizations with a systematic approach to managing and optimizing their observability and testing expenditures by allocating costs across teams, projects, and services. This capability transforms a consolidated bill into detailed breakdowns that facilitate accountability and support internal billing models like chargeback and showback. The platform now extends cost attribution beyond metrics, logs, and traces to include Grafana Cloud Synthetic Monitoring and k6 performance testing, offering deep insights into testing costs and consistent spend attribution across observability and testing workflows. Cost attribution is label-based, allowing users to configure labels such as "team" and "environment" to track usage and expenses, with the data visible in the Cost Management and Billing app. This feature is crucial for centralized observability teams, engineering teams, and FinOps practitioners who require data for cost governance and budgeting. However, it only begins attributing costs from the moment it's configured, and telemetry lacking configured labels will show as unattributed usage. The feature supports up to 1,000 unique label value combinations, and it enables organizations to allocate testing costs using labels appropriate for their projects. As Grafana Cloud continues to develop, it plans to introduce usage attribution for other services, further enhancing cost management capabilities.

Grafana Cloud Frontend Observability offers a comprehensive approach to understanding user experience by going beyond traditional performance monitoring, which focuses primarily on page speed and efficiency. While performance metrics like Largest Contentful Paint and Time to First Byte are important, they often fail to capture user success in completing tasks, revealing the limitations of such metrics in understanding actual user interactions and outcomes. Grafana's tool enhances visibility into user interactions by tracking specific actions, providing geolocation insights, and offering a redesigned errors view that prioritizes the impact of errors on users rather than just their volume. This enables developers to identify and resolve issues that hinder user workflows, such as checkout processes or regional-specific problems, by visualizing user actions and the context around them. By focusing on user actions and geolocation, Grafana Cloud Frontend Observability ensures that developers can detect and address issues more effectively, providing a clearer picture of where users encounter difficulties and how these issues can be resolved to improve overall user experience.

ObservabilityCON 2026, organized by Grafana Labs, is set to take place from October 19-21 at Pier 27 in San Francisco, offering a comprehensive platform for professionals to explore the latest advancements in observability within the AI era. The event will feature a range of hands-on workshops, technical sessions, and networking opportunities, led by Grafana Labs experts and industry leaders, focusing on innovative solutions and strategies for improving AI and observability systems. Keynote speeches from Grafana Labs’ leadership will kick off the conference, followed by detailed sessions on themes like agentic observability, telemetry pipelines, and incident response management. Attendees can also engage with product demos and expert consultations at the Ask the Experts booth. For those unable to attend in person, the conference will extend its reach with ObservabilityCON on the Road, a series of one-day events in São Paulo, London, Madrid, and Bengaluru, providing further opportunities for engagement and learning.

Grafana Labs has been recognized as a Leader in the 2026 Gartner® Magic Quadrant™ for Observability Platforms for the third consecutive year, highlighting its strong vision and execution in addressing the evolving complexities of observability driven by AI advancements. With a focus on reducing operational complexity and enhancing automation, Grafana has introduced tools like Grafana Assistant, an AI-powered feature that aids incident investigation and data analysis, and AI Observability, which monitors AI agents in production. Grafana's Adaptive Telemetry suite helps manage data volumes cost-effectively, and the company's efforts are validated by a high customer satisfaction rating on Gartner Peer Insights™, where Grafana holds a score of 4.5 out of 5 based on 618 reviews. The company's commitment to innovation and user support is further reflected in its expansive user base and the introduction of features that enable teams to customize workflows and gain insights efficiently, fostering a comprehensive observability strategy.

Grafana Assistant revolutionizes data analysis for engineers by integrating over 30 data sources, allowing users to query and visualize information without switching between multiple tools. This AI-powered solution, part of Grafana Labs' commitment to a "big tent" philosophy, eliminates the need to migrate data or standardize on a single vendor by connecting directly to existing data sources. By tapping into platforms like AppDynamics, Azure Monitor, Dynatrace, and more, Grafana Assistant provides a seamless experience where users can ask questions in natural language and receive comprehensive insights that include metrics, logs, and traces all in one place. This capability significantly enhances the efficiency and quality of investigations by correlating data across diverse systems and visualizing it within Grafana dashboards, thus reducing the fragmentation of information and improving the speed of problem-solving.

Modern applications face a variety of potential failures, necessitating a comprehensive reliability testing strategy that involves multiple layers of validation using Grafana Cloud tools. By employing the Swiss Cheese Model, which stacks different testing approaches to catch and address issues at various stages, software reliability can be significantly enhanced. Grafana Cloud offers three main solutions: Synthetic Monitoring, which proactively checks application performance through global probes; Frontend Observability, which provides real-time insights into user experiences and identifies issues not covered by synthetic checks; and k6, a performance testing tool that simulates load conditions to ensure stability under high traffic. An example using the demo web application QuickPizza illustrates how each tool addresses specific gaps, with Synthetic Monitoring detecting latency regressions, Frontend Observability highlighting real user impact, and k6 validating performance under load. Together, these tools allow engineering teams to catch diverse issues, assess their real-world impact, and ensure fixes are effective, ultimately creating a robust reliability testing strategy aligned with the Swiss Cheese Model.

In a recent episode of "Grafana's Big Tent" podcast, hosts Mat Ryer and Tom Wilkie engaged in a discussion with Eric Burns from Anthropic about the company's strategic decision to operate across all major cloud providers, emphasizing the significance of generalizing models to function on diverse hardware. The conversation highlighted the transformative impact of agentic coding tools on software development at Grafana Labs, showcasing how natural language interfaces are reshaping user experience design and enhancing dashboard creation. As Grafana Labs fully embraces AI-driven coding tools like Cursor, Claude Code, and Codex, they recognize a shift towards automated test coverage and seamless system integration, simplifying processes traditionally requiring extensive specification and development time. The episode also touched on personal projects in home automation, illustrating how AI tools expedite coding tasks for non-experts, and concluded with a reflection on the evolving role of engineers in an era where AI increasingly handles code implementation.

Grafana Labs has extended its maintenance commitment for the business intelligence plugins originally developed by Volkov Labs through the end of 2026, following their acquisition. The commitment involves maintaining compatibility with recent Grafana versions, addressing bug fixes, security updates, and managing community contributions. Since assuming responsibility, Grafana Labs has updated the plugins to work with Grafana 13 and React 19, while also engaging with users to understand their needs for these plugins, which offer advanced customization options for specialized visualizations. The company emphasizes transparency in future support plans and clarifies that while they support the plugins themselves, they do not guarantee or troubleshoot user-generated JavaScript code. Grafana Labs encourages feedback from users to help shape the future roadmap and invites communication through various support channels.

Organizations using Grafana Cloud aim to consolidate data from various sources for a comprehensive view of their systems, but this centralization necessitates effective access control to ensure security and scalability. Companies like AcmeCloud, a fictional example, face challenges in manually managing access as they scale, requiring solutions like Single Sign-On (SSO) and System for Cross-domain Identity Management (SCIM) to automate user provisioning. Role-Based Access Control (RBAC) further refines access by assigning roles and permissions based on user functions, while folder permissions and Label-Based Access Control (LBAC) restrict dashboard visibility and data access to maintain tenant isolation. This layered approach allows organizations to efficiently manage user access, balancing simplicity, flexibility, and security in a growing environment, ensuring that while observability is centralized, access remains controlled and secure.