Develocity 2024.3

The Develocity Reporting and Visualization Kit now includes visualizations for test acceleration time-saving realization and potential for Develocity’s Predictive Test Selection and Test Distribution features for Gradle and Maven builds.

Realized test acceleration savings dashboard

This new dashboard facilitates assessing the adoption of Predictive Test Selection and Test Distribution across projects in an organization and observing their effectiveness.

Test acceleration savings dashboard

This new dashboard facilitates assessing the test acceleration that could be achieved from Predictive Test Selection. The measurements are based on simulating the effect of applying Predictive Test Selection to test executions that did not enable it.

Along with projected time savings, the visualization provides several critical metrics to assess prediction accuracy, such as the percentage of predicted task/goal failures and the amount of avoided test class executions.
The visualization will adjust based on the chosen selection profile, helping to find the right balance between time savings and prediction accuracy.

Predictive Test Selection savings potential dashboard

The configuration cache is a Gradle feature that greatly enhances build performance by caching the outcome of the configuration phase and utilizing it for future builds.

A Build Scan now displays the dependency resolution time separately from the configuration cache store time. This helps you better understand the configuration cache’s effectiveness, as dependency resolution typically occurs during execution and isn’t part of the configuration cache overhead.

Configuration cache effectiveness reporting

The /api/builds/{id}/gradle-configuration-cache endpoint of the Develocity API has been extended to include dependency resolution time during storage.

Additionally, Develocity Reporting and Visualization now includes a dashboard to help users analyze the effectiveness of Gradle’s configuration cache, including total configuration cache savings and reasons for configuration misses.
This allows you to identify the most frequent causes of misses, prioritize fixes, and verify that your hit rate does not regress.

Configuration cache dashboard

Develocity Reporting and Visualization includes a new dependency dashboard that offers reporting and search functions for dependency information related to libraries utilized in Gradle and Maven builds.

For example, this feature can be instrumental in identifying projects that use dependencies with known vulnerabilities. The dependency build models can also be accessed through the Develocity API.

Dependencies dashboard

Develocity Reporting and Visualization includes an updated dashboard that facilitates understanding Gradle and Maven plugin usage within your builds.
The filter features enable you to search according to plugin utilization. The dependency build models can also be accessed through the Develocity API.

Plugin dashboard

This release of Develocity introduces several enhancements to build environment resource usage observability, initially released in Develocity 2024.2.  

To help identify which processes within a Build Scan of Gradle, Maven, and sbt consume the most resources during a build, the individual CPU and memory usage of each of the top 5 processes for each metric is now displayed.

The tooltip now also shows the absolute time of the resource usage snapshot for Gradle, Maven, Bazel, and sbt build tools. This helps users to correlate resource usage with other information, such as start times and logs from CI servers.

Furthermore, users can now share specific points on the resource usage graph by clicking a point and sharing the URL, facilitating collaboration.

Resource usage observability

This release delivers significant query performance improvements for users of Develocity Reporting and Visualization backed by AWS Athena. This is achieved via a disabled-by-default option to export data to S3 as compacted ORC files instead of a single JSON file per build. The Athena query engine can query over this data format much more efficiently, significantly reducing query execution times and costs.

Existing configurations using the JSON-based data export are unaffected by this change and will continue to function without interruption upon upgrade.

To use this new capability, please contact your Customer Success manager to discuss the technical requirements for adopting this feature.

Develocity Maven extension 1.22.2 introduces a significant optimization that can reduce time spent on input fingerprinting by up to 40% compared to earlier versions.
The Develocity Maven extension now attempts to reuse checksums from the Maven repository when it is safe to avoid logically redundant operations to determine input identity.
Builds running on ephemeral CI agents are expected to see the greatest performance improvements from this optimization.

Develocity Edge is a new component introduced in this release that allows you to establish Develocity points of presence (or “Edges”) across geographic boundaries. This makes it easier to provide Develocity services closer to where builds are run and makes builds faster by reducing network latency.

In this release, the Develocity Edge can be used as a Build Cache server for Gradle, Maven, Sbt, and Bazel builds. Each Edge automatically manages data sharing with and from the Develocity server and other Edges.

Edge overview

Develocity Edges are assigned to a location. Build users can set their location by logging in to the Develocity application and accessing “My Settings” from the user menu.

My Settings – Location page

The Gradle and Maven Develocity plugins provide a new opt-in “Edge discovery” option, which uses the user’s location setting to automatically assign an Edge to use as a remote build cache. Using an Edge in a nearby location improves build time by reducing network latency and improving throughput.
As each Edge automatically manages data sharing with and from the Develocity server and other Edges, any Edge can be used without sacrificing hit rates.

Previously, using the Develocity build cache node in a distributed fashion required per-user local build configuration changes. The Develocity Edge with “Edge discovery” removes the need for extra local build configuration and allows Edge locations and instances to be updated dynamically as the Develocity server manages all configuration. 

The Edge component is available as a Docker image or standalone fat JAR and is easy to deploy and operate via Kubernetes or any modern computing platform.

The Build Cache administration page now visualizes the following indicators that provide insights into build cache usage patterns and whether the build cache size is undersized, affecting build performance due to a too-short retention period:

• The rate of cache misses for recently evicted entries
• The incoming data storage rate
• For least-recently-used items, how long since the items were last used
• The storage size required to achieve a seven-day retention

Build Cache “Usage” tab

Please note that these metrics apply to the logical build cache served by the Develocity server and Develocity Edges. They do not provide insight into Develocity Build Cache nodes.

Develocity now integrates with npm, the standard package manager for Node.js.

This first release provides basic Build Scan functionality and several build metrics that help you understand, debug, and observe your use of npm.

Build Scan gives you access to this information and more, allowing you to investigate the console output of individual npm invocations.

Build Scan support for npm

The Develocity API now includes a new `npm-attributes` endpoint, which helps you:

• Track how often npm is used
• Identify the machine on which it ran
• Measure how long it took
• Discover the npm and Node.js versions in use across various projects

Develocity now integrates with the Python ecosystem.

The first release includes basic Build Scan functionality and several build metrics that help you understand, debug, and observe the usage of various Python tools.

Build Scan gives you access to this information and more, allowing you to investigate the console output of individual usages of supported tools.

Build Scan support for Python

The Develocity API now includes a new `python-attributes` endpoint, which helps you:

• Track how often Python is used
• Identify the machine on which it ran
• Measure how long it took
• Discover the Python version in use across various projects

Develocity 2024.3 is compatible with PostgreSQL databases up to and including version 17. Previously, version 14 was the maximum supported version. 

The User-managed database migration guide describes the prerequisites for upgrading your Develocity instance’s user-managed database to the latest version of PostgreSQL.

To learn more about running Develocity using a user-managed database, you can read about the trade-offs in either the Kubernetes Helm chart configuration guide or the Standalone Helm chart configuration guide.

The embedded database that Develocity uses, if no user-managed database is configured, continues to run PostgreSQL 14.