April 2014 Summaries

NGINX has quickly become a popular web server, powering a significant portion of the internet's busiest websites, and Datadog is expanding its monitoring capabilities to support NGINX Plus, a commercial version of NGINX introduced in 2013. NGINX Plus offers advanced features such as detailed HTTP request monitoring, health checks for custom pages, and sophisticated load balancing, all of which are now fully supported by Datadog's monitoring platform. Datadog's integration with NGINX Plus includes capturing comprehensive metrics, like connection statuses and response counts, which can be aggregated and visualized for improved infrastructure monitoring and analysis. Tags play a crucial role in Datadog's system, enabling users to manage and analyze data across servers seamlessly, while also allowing for the correlation of events and metrics from various systems. Users interested in exploring these capabilities can access free trials of both NGINX Plus and Datadog, providing an opportunity to enhance their monitoring and insight into HTTP load balancing.

The text discusses the integration of Datadog with Single Sign On (SSO) using Security Assertion Markup Language (SAML). This feature allows users to log in to their Datadog account using their corporate credentials stored in Active Directory, LDAP or other identity stores configured with a SAML Identity Provider. For organizations already using SAML and having an Admin for the Datadog account, configuring this integration involves three simple steps. Once enabled, users can access Datadog through a unique log in URL provided by their organization. The text also mentions that interested users can try out Datadog's performance data analysis feature for free for 14 days.

Datadog has introduced support for Single Sign On (SSO) through its integration with SAML, allowing users to log in to the platform using their corporate credentials stored in Active Directory or other identity stores. This feature is available to organizations already using SAML and Datadog administrators who can configure it in three simple steps. Once enabled, users can access Datadog using a unique login URL that they can bookmark for future use. The company also offers a free trial for 14 days to help users unlock the value of their performance data.

Datadog has released graph markers that allow users to draw regions and lines on dashboard graphs, making it easier to interpret the state of their systems. This feature is useful for highlighting interesting or unexpected conditions without triggering an alert. Users can add markers by opening the graph editor and customizing them using ranges or lines. Multiple thresholds can be set with escalating values to show the severity of a condition. The ability to style, label, and use K/M/G suffixes for big values is also available. This feature aims to make graphs more understandable and can be tried for free for 14 days.

Jonathan Gala from Datadog is excited to announce the release of graph markers, which allow users to draw shaded regions and lines on their dashboard graphs to make it easier to interpret system state. This feature enables users to set boundaries and thresholds for certain metrics and systems, making it clearer what's considered normal and what's not. Adding markers is easy, and users can customize them using ranges or lines, with options to style, label, and use suffixes for big values. The feature also allows users to create multiple thresholds and show progression over time, making graphs even easier to understand.

Jason Moiron, a Datadog software engineer, shares his experience with optimizing the performance of a new service written in Go. He learned that using integer map keys can significantly improve performance, and that AES-NI processor extensions can boost string hash performance. Additionally, he discovered the importance of reducing locking and unlocking in channel communication, which improved throughput by nearly 6x. The author also explored the use of Cgo for optimizing compression, but found that it had some overhead due to coordination with the Go scheduler and memory management. Ultimately, Moiron emphasizes the importance of measuring everything and using the tools available in Go to get the most performance out of the language.

Deploying the Datadog cartridge on OpenShift allows seamless monitoring of OpenShift v2 gears by automatically integrating metrics into Datadog, making it easy to set up graphs and alerts. Although the cartridge is primarily for OpenShift v2, custom metrics can be sent from OpenShift v3 apps using the DogStatsD image. Datadog collects and correlates metrics from OpenShift gears and other infrastructure components, enabling comprehensive monitoring and alerting. Alerts can be configured to notify via email, HipChat, PagerDuty, and other endpoints, and can be set to trigger on specific performance thresholds, like CPU usage. The system is designed to scale with applications, automatically adjusting to monitor aggregate performance across multiple gears. Installation of the Datadog Agent on OpenShift gears requires setting the DATADOG_API_KEY and adding the cartridge to your gear, with further configurations available through editing YAML files. The integration provides a unified view of system performance, helping identify correlations between OpenShift performance and other infrastructure events.

The Datadog's top() family of functions is designed to help users rank, filter, and visualize their performance metrics, allowing them to focus on the most important metrics at any given time. These functions support several ways of ranking time series against each other, such as highest peak values, largest sustained average values, or highest recent values. They also provide flexibility in terms of sorting order and can be used with shortcuts like topN, bottomN, etc. The top_offset function helps skip outliers and concentrate on the next few series for a more granular look into how metric values are distributed across processes.

The Datadog team has introduced the `top()` family of functions to help users cut through metrics clutter and focus on the most important metrics. These functions allow users to rank, filter, and visualize performance metrics in a way that is tailored to their specific needs. By using these functions, users can gain insights into their infrastructure and make data-driven decisions. The `top()` family of functions includes various ranking methods such as maximum, minimum, mean, area, norm, last, and others, which enable users to analyze metrics from different perspectives. Additionally, the `top_offset()` function provides an advanced way to filter out outliers and focus on a specific range of values. These functions are designed to be flexible and customizable, allowing users to create dashboards that meet their unique requirements.