June 2025 Summaries

Grafana is exploring the integration of agentic AI systems into its platform to enhance user experience beyond simple chat responses, focusing on task completion using natural language. This approach involves teaching large language models (LLMs) to "speak" Grafana through the Model Context Protocol (MCP), allowing them to access real-time data and perform tasks such as creating dashboards or managing incidents. At GrafanaCON 2025, the company introduced Grafana Assistant, a tool designed to democratize the observability experience by enabling users to interact with Grafana using natural language, making complex systems more accessible to both seasoned and new users. The development of these agentic systems poses challenges, including creating reliable, modular agents and reducing token noise, but aims to improve productivity and efficiency in observability workflows. Grafana is inviting users to participate in the evolution of these AI-driven workflows by engaging with the open-source Grafana MCP Server and trying the Grafana Assistant in private preview.

Dropbox encountered a significant challenge when a power outage rendered its sole data center hosting Grafana Loki inaccessible, prompting the company to enhance its logging infrastructure. This incident led to the development of a petabyte-scale, multi-region logging platform that can handle up to 6 GB of logs per second with a 30-day retention policy, ensuring availability even during data center failures. The transition involved addressing issues such as high cardinality and memory crashes by imposing strict label control and implementing stream-level controls to prevent one service from overwhelming the system. Gradual deployment and testing were crucial, as was collaboration with Grafana Labs to improve performance with a switch to a Prometheus-style database. This new, robust system has become an integral part of Dropbox's observability stack, allowing the company to phase out its legacy logging system and ensuring high availability and reliability across its operations.

Following a security incident involving a vulnerable GitHub Action, Grafana Labs has implemented Zizmor, an open-source static analysis tool developed by William Woodruff, to enhance the security of their CI/CD pipelines. The vulnerability exploited allowed unauthorized code execution within a trusted environment, but Grafana Labs' proactive security measures quickly mitigated any potential harm. Zizmor, supported by Grafana Labs through a partnership with Woodruff, audits GitHub Actions for risks such as unpinned versions and forbidden uses, helping prevent future security breaches. Despite challenges like GitHub's rate limiting and centralized ruleset configuration, Grafana Labs has integrated Zizmor across over 2,000 repositories using GitHub Organization rulesets to ensure consistent application of security checks. By employing Zizmor in both offline and scheduled modes, they aim to balance comprehensive security assessments with efficient use of resources, preparing them better for any future GitHub Actions vulnerabilities.

Grafana Cloud has introduced several new features and updates to enhance its observability platform, focusing on Kubernetes Monitoring, Fleet Management, and Synthetic Monitoring, among others. In Kubernetes Monitoring, users can now streamline navigation and troubleshooting with new tabs for memory and CPU usage, allowing for more efficient resource management and deeper insights into workload contributions. Fleet Management has been enhanced with the ability to restore pipeline versions and share pipeline components, enabling more dynamic and reusable configurations. Synthetic Monitoring now includes role-based access control (RBAC) for improved security and flexibility in managing workflows. The Adaptive Logs feature has been upgraded to include early detection patterns for better cost management by identifying and filtering out unnecessary telemetry data. Grafana SLO Reports have also been introduced, enabling users to consolidate and analyze multiple service level objectives, while the Grafana IRM mobile app now allows syncing of phone numbers to enhance alert recognition. Additionally, a streamlined Observability menu has been implemented to provide a more unified and efficient user experience, and a new LogicMonitor data source plugin is available for enterprise users to query and visualize device data within Grafana Cloud.

Managing AWS Observability applications in Grafana Cloud using Terraform simplifies configuration and enhances reliability, especially as services scale. By utilizing Terraform, users can define their AWS setups in a clear, recipe-like format that is both trackable and version-controlled, allowing easy rollback to previous states if issues arise. This approach ensures consistent and automatic configurations across environments, minimizing errors and redundancies while promoting collaboration on infrastructure changes. For Grafana Cloud users integrated with AWS, exporting current configurations to Terraform facilitates the management of their setup alongside AWS infrastructure using a declarative method. This method supports version control through Git, automates deployments and updates, and aligns with the DRY principle, thus fostering a robust and efficient monitoring setup. Grafana Cloud offers a user-friendly platform with a forever-free tier and additional plans to accommodate various use cases.

In Morro Bay, California, operational technology specialist Grant Chase revolutionized the city's wastewater treatment operations by developing an award-winning Grafana dashboard, earning him the 2025 Golden Grot Award. Initially faced with outdated technology, Chase utilized Python and Modbus TCP to access raw data and InfluxDB for data storage, then employed Grafana for its visualization capabilities. As the city transitioned to a new wastewater treatment facility, Chase's dashboard evolved into a comprehensive, real-time data command center, offering insights into every process area and integrating compliance monitoring and microorganism video feeds. This innovative dashboard serves as a crucial tool for operators, engineers, and lab staff, exemplifying Grafana's open, community-driven philosophy that fosters creative problem-solving and continuous innovation.

The Factry Historian data source plugin for Grafana enhances the ability of factory teams to derive actionable insights from massive volumes of time-series data generated by industrial systems. By integrating with Grafana, a popular visualization platform known for its support of time-series data, the plugin enables non-technical users to access and analyze data through intuitive and reusable dashboards. This combination addresses the challenge of unlocking industrial data traditionally confined to technical experts, making it accessible to operators, engineers, and decision-makers. The plugin supports hierarchical data browsing, automated metadata handling, and advanced event overlaying, which simplifies the exploration and contextualization of process data, ultimately fostering a culture of learning and improvement within factories. Notably, companies like Agristo have successfully leveraged this tool to streamline operations, enhance troubleshooting, and improve efficiency through self-service analytics, demonstrating the plugin's impact on sustainable and innovative process management.

Grafana Labs has achieved the Google Cloud Ready – BigQuery designation for its Google Cloud BigQuery data source, demonstrating its interoperability and functional compliance with BigQuery. This designation is part of a validation program aimed at enhancing customer confidence in partner integrations. The Google Cloud BigQuery data source for Grafana enables users to query and visualize data from the serverless, AI-ready platform directly within Grafana dashboards, supporting features such as time series and table visualizations, flexible authentication methods, and advanced SQL and visual query editors. The plugin also integrates with the Google Cloud Resource Manager API, allowing access to multiple Google Cloud projects from a single Grafana data source, reinforcing Grafana's commitment to interoperability and flexibility within its observability solutions.

The introduction of Alloy Operator in the latest Kubernetes Monitoring Helm chart version 3.0 aims to simplify and enhance the configuration of telemetry data collection by dynamically setting up necessary collectors based on user choices. The Alloy Operator addresses the cumbersome main values.yaml file of previous Helm chart versions, which often contained unnecessary code, by allowing more streamlined and specific configurations through its dynamic feature influence. This includes the automatic configuration of traditional Alloy objects and the deployment of additional Alloy instances as needed, such as for tail sampling of traces. The Alloy Operator, a Helm-based operator created with the Operator SDK, facilitates easier management of Alloy objects by creating necessary components and performing Helm installs. Despite these changes, best practices in modularity and security are maintained, ensuring that each object remains restricted to required permissions. The transition from previous versions is supported by a migration guide and utility, while feedback from the community continues to drive development improvements.

Grafana Labs is evolving its approach to data source management by refining its "big tent" philosophy, which emphasizes interoperability and vendor neutrality while catering to specific technological environments. Recently, Grafana has deprecated AWS and Microsoft Azure authentication from its original Prometheus data source and introduced dedicated plugins for Amazon Managed Service for Prometheus and Azure Monitor Managed Service for Prometheus. This ensures that users have access to platform-specific features and authentication methods without overcomplicating the core Prometheus functionality, which remains aligned with its open-source roots under the Cloud Native Computing Foundation. With over 100 plugins available, Grafana Labs continues to expand its catalog to support a wide range of licensing models and platform-specific needs, encouraging community and partner engagement to further enhance its offerings. The company remains committed to its core principles of allowing organizations to own their observability strategies and consolidating data from various sources, aiming to be the leading visualization tool for centralized observability.

Grafana Cloud has introduced a new metrics forecasting user interface designed to enhance observability through adaptive alerting by allowing users to proactively address potential issues before they arise. This improved UI streamlines the forecasting process, enabling users to test data for forecastability with live previews and real-time parameter adjustments, reducing guesswork. By integrating forecasting with Grafana Alerting, users can create adaptive alerts that respond to unexpected changes in metrics, such as sales transactions or web traffic, which helps in managing dynamic usage patterns more effectively than static alerts. The system also accommodates daily, weekly, and holiday variations to ensure alerts are triggered only by significant deviations. These enhancements are supported by the augurs library, which performs customization tasks on the frontend, making the process more efficient. Grafana Cloud offers these features free of charge, with options for advanced or enterprise customers to request usage limit increases.

Grafana Labs has released security patches for multiple versions of Grafana, addressing a medium severity vulnerability identified as CVE-2025-3415, which exposed DingDing contact points in Grafana Alerting. This vulnerability, which was reported through a bug bounty program, could allow users with Viewer permissions to see DingDing contact points in plain text due to a configuration oversight, earning a CVSS 3.0 score of 4.3. The affected versions include Grafana 12.0.1 and earlier, and users are advised to upgrade their instances or remove DingDing configurations to mitigate the risk. The timeline of the incident shows a coordinated response involving early notifications to cloud providers like Amazon Managed Grafana and Azure Managed Grafana, ensuring their services remain secure. Grafana Labs encourages users to report security issues through their dedicated page and emphasizes the importance of non-disclosure until a fix is publicly announced. The company also maintains a security section on their blog for ongoing updates and announcements.

Grafana Labs' recent survey on observability trends in Japan reveals that Japanese organizations prioritize cost control and ease of use when selecting observability tools, setting them apart from their Asian and global counterparts. Although Japanese companies are less advanced in adopting observability technologies, with many using three or fewer tools, they may benefit from this cautious approach as the market matures globally. Open source technologies and standards like OpenTelemetry and Prometheus are gaining traction, though Japanese organizations are more in the investigative phase compared to the rest of Asia and the world. The survey highlights a growing interest in AI capabilities, but Japanese respondents remain cautious, with cost and ease of use being the primary selection criteria. The survey was conducted at ObservabilityCON in Tokyo, with participants from various roles and industries, reflecting a broad interest in observability across Japan.

Grafana Tempo 2.8 introduces new features and improvements, particularly in TraceQL, alongside performance enhancements and some breaking changes. Key updates include the addition of the 'most_recent' query hint in TraceQL to return the latest results, the introduction of the 'span:parentID' intrinsic for searching child spans, and new metrics functions like 'sum_over_time', 'topk', and 'bottomk' for better control over data analysis. Performance upgrades focus on reducing compactor memory usage, significantly lowering memory high-water marks by adjusting pooling strategies. Notably, the release fully removes serverless features and changes the default HTTP listen port from 80 to 3200. Looking forward, Grafana Tempo is rolling out new architectures and adding a Tempo MCP server for enhanced query capabilities. Users can engage further with community forums or explore open positions at Grafana Labs for deeper involvement.

Data Points per Minute (DPM) is a key metric in Grafana Cloud that measures how frequently data is sent from each unique time series, acting as a proxy for ingestion volume and system load, and forming a core part of the billing and usage model. While 1 DPM per active time series is standard for most metrics, higher DPM rates can lead to increased costs without necessarily enhancing accuracy or monitoring capabilities, as dashboards and alerts often do not benefit from excessive data. However, higher DPM can be justified in scenarios requiring real-time insights, such as low latency systems or short time window alerts. Managing DPM involves adjusting scrape intervals, identifying high-frequency series with tools like dpm-finder, checking for discarded writes, and using Grafana Cloud's features like Adaptive Metrics and interval processors to optimize data ingestion and reduce unnecessary resource usage. Understanding and controlling DPM helps maintain efficient and cost-effective observability pipelines in Grafana Cloud.

Grafana Play is a publicly accessible sandbox environment hosted on Grafana Cloud, which allows users to explore Grafana features and create dashboards without setting up their own instance. As its popularity grew, Grafana Labs undertook efforts to enhance user privacy by deleting inactive accounts and removing broken or outdated dashboards. They used Grafana k6 for browser testing to identify dashboards with broken panels and implemented scripts to delete over 21,000 inactive users, defined as those not logging in for a year or more and never creating dashboards. The Grafana API, k6 browser module, and other programming tools were utilized to streamline this process, ensuring that user privacy is respected while maintaining the platform's functionality. Grafana Labs invites community contributions to Grafana Play and encourages users to engage in discussions via a dedicated Slack channel.

GrafanaCON 2025, held in Seattle, was the largest gathering of observability enthusiasts to date, featuring a range of technical discussions, user-led sessions, and interactive activities over three days. The event highlighted key developments in Grafana's open-source projects, including the unveiling of Grafana 12 and the introduction of tools like Metrics Drilldown and k6 1.0, enhancing metrics exploration and load testing respectively. Notable sessions included discussions on Grafana Alloy, combining OpenTelemetry and Prometheus for optimized performance, and agent-driven Grafana, showcasing the integration of AI into observability tools. The conference also spotlighted community-driven use cases, such as Firefly Aerospace's successful lunar mission and EA's scalable error monitoring system, illustrating Grafana's role in innovative and large-scale applications. All sessions are now available on-demand, offering insights into the latest in observability strategies and tools.

Firefly Aerospace achieved a historic milestone on March 2, 2025, with the Blue Ghost Mission 1, marking the first fully successful lunar landing by a commercial provider. The mission was supported by meticulous preparation, involving over 500 hours of training and 16 mission simulations, and a mission control center equipped with Grafana dashboards for real-time monitoring. Led by Flight Director Jesus Charles, the team utilized Grafana to maintain situational awareness and make crucial decisions during the mission, including the automated descent and successful touchdown on the moon. The Blue Ghost's journey, which began on January 15, lasted 45 days and saw the team achieve all their scientific goals on the lunar surface, including capturing a solar eclipse. Charles emphasized the importance of ambition, teamwork, and effective tools, encouraging others to pursue their dreams with dedication and perseverance.

Database observability is crucial in modern systems, and OpenTelemetry is enhancing it through semantic conventions, which provide a unified framework for naming and interpreting observability signals across different tools and languages. These conventions ensure consistency by standardizing the way database spans, metrics, and attributes are defined, thus eliminating confusion and improving clarity. Recent efforts have focused on stabilizing OpenTelemetry's database semantic conventions, which now support multiple databases like MariaDB, MySQL, and PostgreSQL, giving developers confidence to build reliable tooling around them. Grafana Labs is actively contributing to this initiative, integrating these standards into their products like Grafana Tempo and Grafana Mimir, to make observability more accessible and consistent. They also offer practical applications of these conventions, such as using OpenTelemetry with a Node.js backend and a Postgres database, enabling easy visualization of telemetry data in Grafana. This approach supports performance tuning and database evaluation by providing valuable insights into query performance and operation types. The ongoing development seeks community input to refine these conventions further, aiming to ease the complexity of instrumentation for developers and platform teams.

ObservabilityCON 2025 is set to take place in London from October 7-9, offering a platform for exploring the latest in observability strategies, featuring keynotes, workshops, and technical sessions focused on innovations like observability as code and AI-powered root cause analysis. The event will also include live demos, expert consultations, and networking opportunities for attendees. Registration opens in July, with early bird pricing available. Grafana Labs encourages community engagement through a call for presentations, inviting submissions until June 27, and offers limited sponsorship opportunities. For those unable to attend in London, a one-day version of the event, ObservabilityCON on the Road, will be hosted in various cities, including São Paulo, Washington, D.C., Dallas, and Madrid, with more locations to be announced.

Synthetic monitoring, specifically through browser checks, is a proactive approach to ensuring the health and performance of web applications by simulating user interactions to catch issues before they affect real users. Grafana Cloud Synthetic Monitoring provides a platform for executing these browser checks, allowing developers to simulate user interactions such as navigation and form submissions across various devices and screen sizes, using the k6 browser testing APIs. This setup involves creating a Grafana Cloud account, navigating to Synthetic Monitoring, and configuring browser checks with specific parameters like script execution, timeout settings, labeling, probe location selection, and test frequency. The results from these checks, which include metrics like uptime and core web vitals, are visualized in a comprehensive dashboard, enabling developers to maintain high-performance user experiences by catching regressions early and monitoring trends in real time. Grafana Cloud offers a user-friendly entry into synthetic monitoring with its free tier, facilitating easy implementation and management of browser checks to ensure reliable service delivery.

Grafana Cloud IRM provides an integrated solution for on-call management and incident response, which can be connected to self-hosted versions of Grafana OSS or Enterprise using a webhook integration. This connection allows users to utilize Grafana Cloud IRM's features, such as on-call scheduling, escalation chains, and incident tracking, without needing to fully migrate to Grafana Cloud. The system supports routing alerts based on custom policies and distinguishes between important and default notifications to tailor responses. Heartbeat monitoring ensures the reliability of the alert-to-incident pipeline by automatically creating incidents if expected alerts are not received within a set timeframe. Users are encouraged to configure their alert rules to ensure only critical alerts trigger a response, and they can take advantage of the free-tier plan in Grafana Cloud to explore these features.

Grafana Labs has integrated the FinOps Open Cost and Usage Specification (FOCUS), an open standard for cloud billing data, into Grafana Cloud to enhance cost management and transparency for users. This integration aligns with Grafana Labs' commitment to open source and community-driven innovation, allowing organizations to analyze cloud costs using standardized, vendor-neutral data. The adoption of FOCUS addresses challenges posed by traditional billing reporting, such as inconsistent metadata and siloed formats, enabling users to make informed financial decisions and optimize observability costs. Supported by the FinOps Foundation, FOCUS provides a consistent language for billing data, contributing to improved efficiency in cost management practices. Grafana Labs plans to launch a redesigned Cost Management and Billing experience later this year, further enhancing cost attribution and spending alerts for users. This initiative is part of Grafana Labs' broader efforts to empower its customers with actionable data and transparency in observability investments.