June 2017 Summaries

InfluxData has reached 300 customers since its release nine months ago, accelerating growth in the Time Series Platform market by delivering fast and accurate data insights to businesses across various industries. The company's modern open source platform is built for metrics and events, empowering developers to build next-generation monitoring and analytics applications with ease. InfluxData has gained significant customer support from prominent companies such as Cisco, Coupa, IBM, LineMetrics, Nordstrom, Tesla, and others, who have chosen the platform for its speed, reliability, and scalability. The company's customers are leveraging InfluxData to gain data-driven insights, improve operational efficiency, and drive digital transformation initiatives.

InfluxQL version 1.3 introduces several new query language features, including mathematical operators such as modulo, bitwise AND, bitwise OR, and bitwise exclusive-OR, which can be used to perform complex calculations on time series data. A new time zone clause allows for the return of the UTC offset for a specified timezone, enabling more precise location-based queries. Additionally, a new duration unit called nanosecond has been introduced, allowing for higher precision in grouping and aggregating data. These features enhance the capabilities of InfluxQL, making it easier to analyze and visualize time series data. The update also includes improvements such as downloads for the TICK-stack, deployment options on cloud or servers, and a call to share user success stories through testimonials.

InfluxData has released new functions in version 1.3 of its InfluxQL, including INTEGRAL() for calculating the area under curves, NON_NEGATIVE_DIFFERENCE() for non-negative differences between subsequent field values, and updates to TOP() and BOTTOM() functions that now support tags as arguments and preserve timestamps when used with GROUP BY time() clauses.

Telegraf is a highly extensible data monitoring system that allows users to collect and store metrics from various sources, thanks to its large collection of plugins, over 300 developed by more than 100 contributors. The system's architecture is based on a configuration-driven agent model, where input and output interfaces are defined using simple Go interfaces, making it easy for developers to contribute and build custom plugins. Telegraf's ease of use and self-documenting features make it an ideal tool for monitoring new projects or existing systems, allowing users to generate config files that showcase the types of metrics they can collect from different sources. The system is actively maintained by a community of contributors who have made significant contributions to its 1.3 release, ensuring ongoing support and development.

The author reflects on the IoT Revolution and its potential to address environmental concerns through smart systems that learn and adapt, such as self-driving cars and smart homes. While acknowledging past fears about technology's ability to solve problems, the author notes that recent advancements in machine learning and sensorization have brought significant promise to sustainability and efficiency gains. The author's company, InfluxData, is working to support this future by building an open-source platform for data processing and analysis, enabling customers like BBOXX and Spiio to create "smarter" systems that optimize resource use and minimize environmental impact. Despite lingering concerns about humanity's ability to manage resources sustainably, the author expresses optimism about the potential of technology to drive positive change.

Draft for Kubernetes is a tool that helps developers quickly prototype microservices and expose them on a publicly available domain. It utilizes Docker, Helm, and Kubernetes Ingress Controllers to make applications publicly available in a hurry. Draft generates a generic Dockerfile and helm chart for the application, creating an ingress resource to expose it at a URL. To use Draft, you need a Kubernetes cluster, a domain name, an account on quay.io or hub.docker, a working Helm installation, and a golang development environment. The tool supports basic python, ruby, php, node, and java applications. After installing the dependencies and creating a sample application, users can run the `draft init` command to set up Draft, which spins up a deployment and some draft pods in the kube-system namespace. Once set up, users can use Draft to create their application by running the `draft create` command, which scaffolds out the helm chart and Dockerfile, as well as creates a draft.toml file that controls different environments for the application. The final step is to run `draft up`, which builds the Docker image, pushes it to the configured container registry, runs Helm install on the preconfigured helm chart, and deploys the application to Kubernetes. Draft provides a way to quickly iterate on things like webhooks or APIs by rebuilding the image and pushing changes live with a single command.

The conference focused on monitoring and its evolution, with a major theme being the idea that "metrics are dead" and tracing and events are becoming more prominent. However, the author argues that metrics will never die due to their importance in summarizing and analyzing data at scale, particularly when dealing with large volumes of data. Metrics provide a method for summarization, allowing operators to visualize or analyze data effectively, and are crucial even in moderately sized data sets. The conference also highlighted the importance of logs, tracing, and events in monitoring, with tracing being a hot topic but ultimately limited to distributed/microservices environments where metrics, events, and logs remain more relevant.

Chronograf, a time series database management tool, allows users to manage alert history by storing data in the `chronograf` database and `alerts` measurement in InfluxDB. By default, this data is subject to an infinite retention policy but can be expired by shortening the duration of that policy through the Admin page in Chronograf. The recommended minimum duration for expiring alert history data is one hour. Additionally, starting with version 1.3.2.0, Chronograf's Data Explorer supports writing data to InfluxDB databases, allowing users to write data directly within the tool instead of using the deprecated web admin interface.