December 2025 Summaries

Joining Grafana Labs as a senior solutions engineer, Julius Vogt faced the challenges of entering a new industry and dealing with imposter syndrome due to his unfamiliarity with observability concepts. He describes the importance of shortening the "inner loop," the learning process involving trial and error, by utilizing Grafana Assistant, an AI tool designed to aid users in navigating Grafana's complexities. This AI assistant functions like a knowledgeable pair programmer, offering guidance through contextual answers and draft queries, thus connecting theoretical knowledge to practical application and accelerating the learning curve. Vogt emphasizes treating the AI as a teammate by providing clear context to get precise assistance and highlights its role in onboarding by explaining label choices, rate windows, and avoiding common pitfalls. He encourages experimenting by comparing traditional methods with AI assistance to experience a more efficient and less frustrating learning path, ultimately aiding in staying productive and engaged. Grafana Cloud is presented as an accessible platform for managing metrics, logs, and traces, with a free tier available for new users.

As Ocado Technology’s grocery platform grew to include 12 retail partners worldwide, its observability system became overly complex and costly, utilizing 13 tools across 7 vendors. This fragmented setup led to operational inefficiencies, requiring engineers to frequently switch tools and interfaces to troubleshoot issues, which slowed down response times and increased costs. In response, Ocado consolidated its observability tools onto Grafana Cloud, choosing it for its open-source standards and cost-efficient governance features. By migrating from New Relic to Grafana Cloud, Ocado streamlined its architecture, improved cost savings, and enhanced troubleshooting speed with centralized logs, metrics, and traces. The transition involved using OpenTelemetry and Micrometer for metrics instrumentation and Fluent Bit for logs, which facilitated a swift adoption process with most applications migrating within two weeks. Despite initial challenges, the migration empowered Ocado’s engineers with a unified toolkit, enabling proactive monitoring and future-proofing their observability strategy. Ocado continues to explore further optimizations, such as integrating incident response into Grafana Cloud, marking the migration as a strategic move towards operational efficiency and scalability.

High cardinality in a Prometheus environment refers to the situation where a database has a vast number of unique time series, often resulting from metrics with quickly changing label values. This can lead to increased resource usage, higher costs, and degraded performance of both the database and the applications generating the metrics. High cardinality can negatively impact applications by consuming excessive CPU and memory resources, slowing down operations, and causing errors in queries and Grafana dashboards due to data limits. While labels are essential for telemetry, they should be used judiciously to avoid these issues, and alternative observability pillars like logs and traces can provide necessary information without increasing cardinality. Developers are encouraged to review and optimize their metric labels to prevent high cardinality and collaborate with their Prometheus management teams to address any existing issues. Tools like Grafana Cloud's Adaptive Metrics can help manage high cardinality by aggregating series to alleviate the load on the system.

HTTP checks in Grafana Cloud Synthetic Monitoring offer a proactive approach to ensuring the reliability of applications by monitoring HTTP endpoints for availability, responsiveness, and correct status codes. By setting up HTTP checks, users can track response times, capture errors, and validate returned status codes from multiple global locations, thereby detecting and addressing issues before they affect end users. The process of creating an HTTP check involves configuring various settings such as request type, uptime criteria, probe locations, and test frequency, and setting up alerts to notify teams of performance deviations. Grafana Cloud provides a preconfigured dashboard for visualizing check results, including metrics like uptime, latency, and SSL expiry, facilitating continuous performance monitoring. This functionality allows for comprehensive, real-time monitoring, enabling teams to maintain high service reliability and user satisfaction.

In 2025, Grafana dashboards showcased their versatility across a wide range of innovative applications, both on Earth and beyond. Notably, Firefly Aerospace used Grafana for lunar landing telemetry during their Blue Ghost Mission 1, while Pixxel monitored Earth's health through satellite data. On the ground, Grafana helped optimize wind turbine performance in Spain, track environmental impacts of cloud computing in the UK, and monitor the longest battery-powered train journey. Unique personal projects included a Star Wars-themed dashboard, Taylor Swift career analytics, and a college dorm laundry tracker. The Golden Grot Awards highlighted outstanding dashboards like those for wastewater management and the International Space Station. Grafana's adaptability was further exemplified by creative uses in tracking March Madness, houseplant health, autonomous boat journeys, and more, illustrating the tool's broad appeal and impact on decision-making across diverse domains.

In an episode of the "Grafana’s Big Tent" podcast, hosts Mat Ryer and Tom Wilkie discuss the evolution of Prometheus with Julius Volz, its co-creator, and Richard Hartmann, a Prometheus maintainer. Prometheus, originally developed at SoundCloud as an internal tool to address monitoring issues, has grown into a significant open-source project in the observability domain. The podcast highlights Prometheus's journey from version 1.x to 3.0, emphasizing its text-based metrics format, which enhances accessibility and performance. Prometheus's compatibility with Kubernetes, despite not sharing origins, was quickly established due to shared inspirations from Google's tools. Prometheus 3.0 is noted for its performance improvements and the introduction of native histograms, which provide more detailed insights into data. The episode also discusses the governance changes inspired by OpenTelemetry to broaden the contributor base. The podcast encourages engagement from listeners with stories or feedback.

Grafana, traditionally known for visualizing time series metrics, has been utilized to create an interactive dashboard for survey data, enhancing user engagement and data analysis. The process involved collecting over 1,250 survey responses via Google Forms, processing them in Google Sheets, and then importing the data into BigQuery to facilitate SQL-based transformations. The project highlighted the challenges of data transformations and the benefits of using SQL expressions and dashboard variables to filter data by demographics like region and role. Visualizations included standard options like bar and pie charts, as well as complex representations using the Business Charts panel powered by Apache ECharts. This approach enables users to analyze business metrics in a visually appealing manner, providing insights into industry trends and allowing for personalized analysis. As part of the ongoing Observability Survey, the dashboard serves as a valuable tool for both the community and industry observers to assess the state of observability, with future surveys promising more comprehensive visualizations.

Adaptive Traces, part of Grafana Cloud's Adaptive Telemetry suite, introduces a refined approach to tracing in distributed systems by utilizing tail sampling to retain only high-value traces, such as those with errors or elevated latency, thus reducing data storage costs and noise while preserving critical insights. Unlike head sampling, which can miss important traces by making sampling decisions early, tail sampling waits until a trace has collected most of its spans before determining whether to keep or discard it, allowing for more precise capture of performance issues. This method helps teams manage costs and complexity without the need for a self-hosted sampling pipeline. Adaptive Traces also addresses the challenge of skewed metrics by generating accurate metrics from raw trace data before any downsampling occurs, ensuring reliability. It integrates with OpenTelemetry and provides users with flexible policy configurations to tailor their tracing needs, supporting anomaly detection and seamless adaptation within Grafana Cloud's infrastructure. This capability is available across all Grafana Cloud tiers, including the free tier, and is part of a broader suite that includes Adaptive Metrics, Logs, and Profiles to enhance observability and performance analysis.

OpenTelemetry, a vendor-neutral open standard for application instrumentation, offers two primary approaches: eBPF instrumentation and runtime-specific SDKs. eBPF, a Linux kernel feature, provides comprehensive baseline metrics by operating at the network level, offering benefits such as universal applicability across programming languages and legacy systems, and the ability to generate service graph metrics out of the box. SDKs, on the other hand, deliver deep transaction insights and distributed tracing, enriching application logs with trace contexts and offering runtime-specific metrics. A hybrid model that combines eBPF for baseline metrics and SDKs for detailed application-level insights is recommended, as it ensures broad observability coverage with minimal overhead. This approach is supported by tools like Grafana Beyla, which work alongside SDKs to avoid signal duplication, and is integrated with platforms like Grafana Cloud to facilitate easy implementation. Despite some limitations, such as current Linux-only support for eBPF, efforts are underway to enhance cross-platform compatibility and tighter integration between eBPF and SDKs for a more unified observability solution.

Instrumentation Hub, introduced by Grafana Cloud, offers a streamlined and scalable approach to observability in complex, fast-moving environments by providing a control plane for remote discovery and selective auto-instrumentation. This tool addresses the challenges of traditional manual setup and indiscriminate data collection by allowing teams to strategically apply instrumentation only where necessary, thus maintaining data quality and controlling costs. Instrumentation Hub is built around the concept of a hierarchy of observability needs, starting with infrastructure visibility and moving towards custom instrumentation, ensuring a structured adoption without overwhelming complexity. It integrates Grafana Alloy and OpenTelemetry Collector, facilitating immediate insights through guided workflows for Kubernetes Monitoring and Application Observability. The service allows continuous control over observability settings, enabling users to adjust data collection in response to changing needs, and aims to expand its capabilities for deeper insights and technology-specific integrations. Grafana Cloud supports this initiative with an accessible platform that includes a free tier, making it easy for organizations to begin their observability journey.

In 2025, Grafana Labs made significant strides in integrating AI into its products, focusing on creating genuinely useful tools rather than jumping on the AI bandwagon. Grafana Assistant, an AI agent developed from a hackathon idea to general availability within nine months, became a standout feature, enhancing user experiences by simplifying complex tasks like building dashboards, investigating issues, and understanding telemetry data using natural language. The company emphasized building AI solutions that prioritize human-in-the-loop interactions, ensuring users remain in control while benefiting from AI's assistive capabilities. Grafana Labs also developed AI observability tools for monitoring large language models (LLMs), further cementing their position as leaders in the AI space within the observability market. Their approach leverages years of open source development, allowing for rapid AI advancements without the need for extensive data training. The company plans to expand Grafana Assistant's capabilities and integrations, underscoring a commitment to improving observability practices through AI innovations.

Integrating Grafana-managed alert rules with ServiceNow through ServiceNow's scripted REST API allows for the automatic capture and processing of alerts in ServiceNow’s events table, enhancing incident workflows, escalations, and ticket creation. This integration enables the parsing of Grafana's JSON alert payloads and the insertion of data into ServiceNow tables, preserving all alert details and facilitating downstream automation within ServiceNow’s incident management ecosystem. The scripted REST API is particularly effective in handling Grafana's complex payloads, which often include nested arrays and multiple instances, by disassembling these payloads and mapping custom labels into the correct ServiceNow table fields. The process involves setting up a ServiceNow table, creating a scripted REST API to manage the data flow, configuring Grafana's contact point to alert ServiceNow, and conducting tests to ensure seamless integration. This setup helps bridge the gap between Grafana’s flexible alert payloads and ServiceNow’s structured data management, allowing platform engineers, SREs, and solutions engineers to efficiently automate incident workflows.

Martin Falch of CSS Electronics explores the integration of AI with Grafana Assistant for analyzing and visualizing CAN bus data, which is vital for industries like automotive and industrial manufacturing. CSS Electronics provides CAN bus data loggers that collect vast amounts of data, which can be challenging to analyze and visualize, especially for users without a data science background. Grafana Assistant simplifies this by allowing users to perform complex data analyses and generate customized dashboards using natural language prompts, without needing extensive setup or coding skills. Unlike ChatGPT, which has limitations in file size and data sharing, Grafana Assistant can handle large data lakes and allows easy exploration and visualization of data through customizable dashboards. The tool's potential is highlighted by its ability to create dashboards based on high-level or detailed prompts, offering a significant time-saving advantage. Despite some limitations, such as occasional query errors and a lack of script execution, the assistant is seen as a promising tool for enhancing data visualization and exploration capabilities.

Season 3 of "Grafana's Big Tent" podcast kicks off with a lively episode recorded at GrafanaCON 2025 in Seattle, featuring hosts Mat Ryer and Tom Wilkie from Grafana Labs, along with guests Ivana Huckova, Andrew McCalip, and Brad Fitzpatrick. The episode delves into various topics such as the quirks of home labs, successful and failed IoT projects, and the intricacies of 3D printing and CNC machining. Discussions include Ivana's adventures in plant and candle monitoring, Andrew's autonomous drone-ship, and Brad's practical 3D printing endeavors. The conversation touches on the challenges and humorous mishaps faced in tech tinkering, including the dangers of maritime practices and the philosophical debates around redundancy in ocean-going drones. The podcast continues to embody Grafana's "big tent" philosophy, encouraging open dialogue and innovation within the observability community, and invites listeners to share their own stories or feedback.

In 2025, Grafana Labs marked a year of substantial innovation and growth, highlighted by the release of Grafana 12 and significant milestones in open-source projects like Mimir, k6, Beyla, Faro, and Alloy. The company launched AI-powered tools such as Grafana Assistant and the Adaptive Telemetry suite, enhancing the observability experience by simplifying data interaction. Grafana Labs also expanded its global community, notably establishing a new subsidiary in Japan, and was recognized as a Leader in the Gartner Magic Quadrant for Observability Platforms. The third annual Observability Survey revealed growing industry interest despite challenges in cost and complexity. Financially, the company achieved over $400 million in annual recurring revenue, reflecting its expanding influence in the observability sector. CEO Raj Dutt emphasized the importance of the open-source community in reaching these milestones and expressed optimism for continued innovation and collaboration.

Grafana has decided to deprecate the Drilldown Investigations feature from its open-source software (OSS) and deprovision it from Grafana Cloud, as it never achieved general availability and faced low adoption compared to core Drilldown applications. This decision will not affect the Drilldown apps for metrics, logs, traces, profiles, or the classic Explore feature. The Drilldown Investigations aimed to consolidate findings from different signals to expedite incident response, but its additional complexity did not offer substantial day-to-day benefits for users. Users are encouraged to convert existing investigations into dashboards or copy important notes before the feature's removal in early January. Grafana Assistant Investigations remains available as an alternative for incident analysis, with integration into Grafana's AI-powered chat bot for follow-up actions.

Cloudflare Workers, a platform for deploying serverless functions globally, has integrated with Grafana Cloud to enhance observability by exporting OpenTelemetry logs and traces directly to Grafana Cloud. This integration employs the OpenTelemetry Protocol (OTLP) for seamless data streaming, allowing developers to visualize and analyze application performance through pre-built Grafana dashboards. These dashboards enable detailed insights into request patterns, regional performance, client environments, and execution contexts, aiding in debugging, capacity planning, security monitoring, and user experience optimization. The system uses a push-based pipeline for telemetry data, requiring minimal configuration without the need for additional infrastructure. This integration empowers developers to efficiently track and resolve issues, optimize resources, and improve application reliability across a distributed global network.

The Grafana Image Renderer, a backend service for rendering Grafana dashboards and panels as images, has undergone significant improvements with the release of version 5.0. This update enhances the service's performance, reliability, and security, including a rewrite of the service, better heuristics for higher-quality image rendering, and significant security enhancements for Grafana Cloud users. The new version has moved to a Go-based implementation, resulting in reduced memory usage and allowing for low-level security improvements, such as isolating the Chromium browser within a custom sandbox. The service now offers extensive configuration options, including rate-limiting based on available memory and acceptance tests to ensure changes do not disrupt functionality. The plugin version of the Image Renderer has been deprecated, requiring users to deploy it as a separate service. With these updates, Grafana Cloud users automatically receive the benefits, while self-managed environments must migrate to the new configuration to take advantage of the improvements.

Grafana Labs introduces Interactive Learning, an innovative feature embedded into the Grafana platform, designed to provide contextual, in-product guidance, eliminating the need to search through external documentation. This feature, available in all editions starting with Grafana 12.3, leverages Grafana’s runtime and DOM to offer dynamic, tailored learning resources directly within the user interface. Interactive Learning includes context-aware documentation, videos, step-by-step examples, and interactive guides that simplify tasks by highlighting UI elements or automating actions. The system operates offline by default to ensure privacy but can use a recommender service for more personalized assistance. Future developments include custom interactive guides and integration with Grafana Assistant for enhanced data exploration. Users can explore Interactive Learning in Grafana Play, with the feature being a collaborative, open-source project shaped by community feedback.

Grafana Cloud's Service Center is a newly introduced feature designed to enhance service reliability and operational culture within engineering organizations by providing a centralized hub for service-related activities. It aims to improve visibility into service performance and interdependencies, helping teams manage on-call rotations, tackle technical debt, and prioritize service reliability. By consolidating data from various tools and dashboards, Service Center facilitates operational reviews that translate data into actionable insights, reducing manual efforts and preventing engineer burnout. It offers features such as dedicated service pages with performance metrics, alerts, incidents, and on-call information, allowing teams to quickly assess service health and make informed decisions. Additionally, the Service Center empowers stakeholders with direct access to essential service performance data, eliminating the need for manual data compilation and enabling engineers to focus on development and issue resolution. Available for free to all Grafana Cloud customers, Service Center requires proper labeling of SLOs, alerts, and incidents for seamless integration and offers an easy way to navigate through observability services with static metadata links.