September 2021 Summaries

Lightbend, a leader in cloud-native applications, uses Grafana Cloud to enhance the monitoring of its platform-as-a-service, Akka Serverless, which co-locates data and functions for cloud-native applications running on Kubernetes. By leveraging Grafana Cloud, Lightbend improves observability, allowing the team to understand system performance from an external perspective and respond quickly to meet service level agreements. Key dashboards in Grafana include ones for operations and synthetic monitoring, which save time for support teams by providing quick insights into production deployments and resource consumption. The Akka Serverless metrics dashboard and the Kubernetes dashboard are particularly valuable for monitoring high-level metrics and visualizing Kubernetes clusters, respectively. Looking forward, Lightbend aims to integrate Grafana Cloud Logs and Loki to correlate metrics and logs seamlessly, demonstrating how advances in technology have transformed monitoring and observability for managing large-scale applications.

The U.S. Department of Defense (DoD) has formally authorized the use of Grafana, Grafana Enterprise, and Loki, along with Prometheus, for its 100,000 developers, marking a significant step in modernizing its software development practices. Previously reliant on outdated methods, the DoD launched the Enterprise DevSecOps Initiative in 2018 to accelerate the delivery of secure and resilient applications, particularly in the fields of AI, machine learning, and cybersecurity. This initiative includes utilizing Kubernetes and other open-source technologies, as well as establishing Iron Bank, a repository of vetted and secure container images. The authorization allows DoD teams to deploy Grafana's software instantly, enhancing their ability to build critical applications efficiently with essential observability tools that provide vital insights and support a wide range of data sources, crucial for an organization of the DoD's size and complexity.

The Salesforce plugin for Grafana allows users with a Grafana Cloud account or Grafana Enterprise license to visualize Salesforce data (SFDC) on Grafana dashboards, enabling the correlation of sales data with other system metrics and logs. This integration enhances observability by allowing companies to track their sales pipeline performance in relation to system events, potentially impacting revenue generation. Grafana provides three methods to query Salesforce data: through native Salesforce Object Query Language (SOQL), a query builder for users unfamiliar with SOQL, and existing Salesforce reports. Users can create dashboards that combine Salesforce data with metrics from other data sources, such as Prometheus, to gain insights into how system performance affects sales outcomes. The plugin facilitates adding annotations to charts for richer data interpretation, thereby helping companies understand the impact of system outages or software releases on sales pipelines.

Dave Schmid shares his journey of using Grafana to visualize real-time data from the WeatherFlow Tempest, an IoT smart home weather station. Initially inspired by his passion for weather stations, Schmid explores various ways to collect and visualize weather data, starting with using Loki to log metrics and later transitioning to InfluxDB for more complex data handling. His project expands to include multiple dashboards that track not only current weather conditions but also forecasts, using APIs from both local networks and the WeatherFlow Cloud. Schmid also builds observability frameworks into his project to understand performance metrics and optimize his system, especially as he integrates feedback and needs from the WeatherFlow community. This leads him to create more user-friendly and flexible solutions, such as a standalone WeatherFlow Collector and a WeatherFlow Dashboard AIO, to help other users, regardless of their technical experience, visualize their data efficiently. Throughout, Schmid emphasizes the importance of using Grafana's flexibility to access diverse data sources and create engaging and informative dashboards.

Distributed tracing is an essential practice for understanding the complexities of modern distributed systems, particularly as traditional monolithic architectures evolve into microservices. Observability, characterized by the three pillars of metrics, logs, and traces, allows teams to inspect and analyze system behavior effectively. The focus on distributed tracing addresses the challenge of coherence loss in microservices by instrumenting systems to capture runtime information, including variables and stack traces, as timestamped data. This process, while possibly adding performance overhead, greatly aids in analyzing system health, diagnosing errors, and identifying performance bottlenecks. The implementation of distributed tracing, such as through the OpenTelemetry project, involves setting up spans and traces that illustrate the causal relationships within system requests. The article exemplifies this with a demo using JavaScript microservices orchestrated by Docker and visualized through Grafana Cloud, demonstrating how distributed tracing can provide actionable insights into system performance and reliability.

Sitech Services, a Dutch technology firm, is leveraging Grafana Cloud to modernize industrial IoT monitoring solutions at Chemelot, a major industrial park in the Netherlands. By integrating Grafana's capabilities, Sitech's Asset Health Center is able to analyze and unify data for clients, enhancing safety and process efficiency across chemical and process industries. Grafana's dashboards and tools, such as state timelines and flowcharting, enable Sitech to provide preventive and predictive maintenance, ensuring optimal equipment health and longevity. The firm transitioned from an older self-managed solution to Grafana Cloud and Grafana Enterprise, allowing them to focus on core competencies while benefiting from remote monitoring, user customization, and improved user experience. This shift has made it possible for Sitech to monitor hundreds of assets remotely and develop new integrated solutions to further improve industrial processes.

CSS Electronics, co-led by Martin Falch, utilizes Grafana and InfluxDB to streamline the visualization of CAN bus IoT data for monitoring vehicles and machinery across various industries, including automotive and maritime. The integration process involved addressing challenges such as decoding raw CAN data, storing binary log files, and ensuring user-friendly setup without central hosting. By leveraging Python scripts, the CANedge data loggers' output is processed and pushed to InfluxDB, making it accessible through Grafana dashboards that offer customizable and visually appealing solutions. The integration has enabled companies like Kverneland and HAVELSAN to perform remote diagnostics and real-time data analysis, significantly enhancing operational efficiency. Looking forward, CSS Electronics aims to expand its integration capabilities with Grafana's real-time streaming features to improve equipment diagnostics and data management further.

The article explores the use of the xk6-kafka extension with k6, a versatile open-source load testing tool, to test Apache Kafka, a powerful event streaming platform. k6 extensions, written in Go, expand k6's capabilities to cover various protocols and scenarios. The xk6-kafka extension provides APIs for producing and consuming messages, creating topics, and managing Kafka servers, offering a streamlined way to simulate high message volumes and assess system performance. The guide outlines installation steps, including building k6 with the Kafka extension and running Docker for Kafka setup, and provides code examples to test Kafka's performance by simulating producer and consumer applications. It highlights k6's ability to create custom metrics for detailed performance analysis and mentions the availability of pre-built test scripts and community support for further exploration. The article concludes by emphasizing k6's enhanced ability to facilitate scalable load testing of Kafka, suggesting Grafana Cloud as a starting point for integrating these tools into performance testing workflows.

To effectively monitor the popularity and usage of numerous GitHub repositories, a system using Grafana and Prometheus was developed to automate the collection and visualization of traffic data. This GitHub Traffic tool leverages GitHub's REST APIs to gather information such as views, clones, and stars from all repositories within an organization, transforming this data into Prometheus metrics. These metrics are then visualized using Grafana, providing a streamlined and scalable solution compared to manually collecting data for each repository. The tool's implementation is open-source, allowing others to deploy it by setting up an environment file and running a Docker container. Although the initial version focuses on a limited dataset, future iterations may expand its capabilities. The project encourages community feedback and further engagement through Grafana's Community Slack and GitHub Repo issues.

Jita Chatterjee is a Senior Technical Writer at Grafana Labs, focusing on enhancing the open-source documentation for Grafana, driven by the docs-as-code practice and the potential to boost community adoption. Based in Lexington, Massachusetts, she is new to the open-source domain but is eager to contribute to more projects. Outside of work, Jita enjoys reading travel blogs, visiting various places, and experimenting with international recipes at home, with a particular fondness for fish curry and rice, reflecting her Bengali heritage. Her professional motivation comes from the challenge of simplifying complex technical features for her audience, ensuring users benefit from her clear explanations.

In the upcoming v2.4 release of Loki, the requirement for log entries to arrive in order by timestamp will be removed, addressing one of the most-requested features. This change allows out-of-order writes to be accepted, simplifying setups where multiple instances of an app or complex systems result in logs arriving slightly out of order, which previously led to data being dropped. The new feature can be configured easily, with a negligible performance cost, and applies either to an entire Loki cluster or on a tenant-by-tenant basis. Users of Grafana Cloud, a managed platform that integrates metrics, logs, traces, and dashboards, can already benefit from this functionality, as out-of-order writes up to one hour old are accepted. This advancement aims to accommodate various logging scenarios, particularly in high-load systems or those using ephemeral jobs like Lambdas, where maintaining strict log ordering is challenging.

Grafana Cloud introduced several new features and updates in September 2021, enhancing user experience and data management capabilities. Noteworthy additions include new visualization options such as the Geomap and Annotations panels in Grafana 8.1, improved time series visualizations, and a "prettify JSON" option for easier log analysis. These updates are complemented by enhanced synthetic monitoring, customizable reporting, and query caching to reduce load times and costs. Users can also benefit from integrations with services like Kafka, Jenkins, and GitHub, allowing for quicker deployment of preconfigured dashboards. Additionally, enterprise-grade features have been expanded with the introduction of OAuth configurations and advanced reporting options. The expansion of Grafana Cloud to the EU region offers users the ability to select where their data is stored, while new commands for cortex-tools help optimize metric usage. These updates are designed to streamline data observation and enhance the overall efficiency of monitoring workflows on Grafana Cloud.

A live webinar titled "Intro to load testing with Grafana and k6" is scheduled for September 16, highlighting the use of observability tools and load testing to assess application performance before deployment. Developer Advocate Nicole van der Hoeven will provide an overview of k6, a load testing platform tailored for modern engineering teams, and demonstrate how to quickly create realistic load testing scripts to simulate production environments. The session will cover integration options with Grafana and Prometheus to allow teams to visualize load test results within familiar interfaces. Participants will learn to run comprehensive thousand-user load tests across various geographical regions in the cloud and use Grafana to interpret the data effectively. The webinar also offers insights into how Grafana Labs employs k6 internally, promising attendees an understanding of practical applications and benefits.

Integrating load testing into a CI/CD pipeline using AWS CodeBuild and Grafana's k6 involves setting up automated performance tests to evaluate system performance under specific loads. k6, an open-source tool, is used to create load tests that simulate virtual users and measure performance thresholds, such as response times. AWS CodeBuild, a managed continuous integration service, automates the execution of these tests as part of the build process. The guide explains how to configure these tests using Docker and buildspec.yml files, and details the process for running tests both locally and in the cloud, leveraging AWS services like Amazon EventBridge for scheduling. This integration allows for early detection of performance regressions, ensuring system reliability before deployment.

AWS IoT SiteWise Edge, in conjunction with Grafana, offers a powerful solution for industrial customers to collect, process, and monitor industrial data both in the cloud and on-premises. The integration allows users to visualize industrial equipment data through Grafana dashboards, facilitating real-time monitoring even when connectivity is intermittent. SiteWise Edge enables data to be processed locally, which is crucial for industries like manufacturing and oil & gas that often face connectivity challenges. By utilizing asset models, context is added to raw data, which is then transformed into actionable business metrics. Customers can use AWS IoT SiteWise's console to configure these models and visualize data through either SiteWise Monitor or Grafana, with the latter offering flexibility to integrate data from multiple sources into a single dashboard. The system also supports routing data to other AWS services for extended storage, reporting, and machine learning applications, ensuring comprehensive data management and operational efficiency improvements.

The Lightstep Metrics plugin for Grafana, introduced by software engineer Chris Sackes, allows users to integrate Lightstep's metrics reporting and analysis capabilities directly into their Grafana instances. By creating a data source plugin, the integration enables users to view metrics data reported to Lightstep and investigate changes in system behavior via Grafana's visualization tools. The plugin leverages Grafana's resources for data source plugin development and requires user configuration of organizational details and an API key, secured with Grafana's secureJsonData mechanism. The data fetched from Lightstep is transformed into WideDataFrames for enhanced performance, allowing for efficient chart rendering in Grafana. Additionally, the plugin includes a feature that links Grafana charts to Lightstep's Change Intelligence, enabling users to seamlessly explore system changes directly through the Lightstep platform.

Grafana Cloud has introduced query caching, a feature designed to improve performance and reduce costs for popular dashboards by storing query results for faster retrieval, thereby decreasing load times and API request costs. This feature is particularly beneficial during high-traffic events, such as Black Friday and Cyber Monday, by lowering total wait times and operational costs significantly. To enable query caching, users must activate it in the data source settings, with support available for various data sources including CloudWatch, Google Cloud Monitoring, and MySQL, among others. Once enabled, users can adjust the Time to Live (TTL) settings to control data refresh rates, with the default set to 5 minutes, while specific dashboard panels can be set to refresh more frequently if needed. Query caching is accessible to all Grafana Cloud Pro and Advanced users, with a 14-day free trial available for new users, and further information and support can be accessed through the Grafana Cloud Community Slack.

Grafana ObservabilityCON 2021 is set to take place online from November 8-10, following the success of the previous year's virtual event that attracted over 9,000 participants globally. The event aims to bring together observability practitioners to share insights, success stories, and developments in the field, with highlights including game-changing announcements from Grafana Labs, such as Grafana Loki 2.0 and Grafana 7.3's exemplar support. The conference will feature roadmap updates, demos, and workshops on metrics, logs, tracing, and alerting, as well as talks from the community on various observability topics. Grafana Labs encourages participation from first-time speakers and individuals from underrepresented groups, with a call for papers that closed on September 22. Registration for the event is free, and attendees can look forward to updates on the speaker lineup, schedule, and workshop availability.