Stackify Changes Pricing Model for Retrace

Stackify has changed the pricing model for its APM tool, Retrace. This change in pricing is in line with usage-based shifts in technology along with changing architectures.

Earlier this year, software vendor Stackify began phasing in a new pricing model for Retrace, its application performance monitoring (APM) tool.

Unlike most traditional APM tools, such as Dynatrace and AppDynamics, which tend to be more focused on the needs of infrastructure and operations teams, Retrace is an APM tool marketed primarily to developers as a way to improve performance and quality of their code.

Retrace is a Software as a Service (SaaS) solution. Users install the Retrace agent on their physical or virtual machines to collect log data, errors, and exceptions. These agents send the data to Retrace, where it is centralized for analysis and reporting to generate actionable insights. Retrace users can log in through their browser to analyze performance using the graphical user interface (GUI).

In its older pricing model, Stackify charged Retrace users in a very traditional way: based on the number of agents deployed across their infrastructure. Each physical or virtual machine would require an agent for visibility into the performance of applications running on that box.

In the new model, Retrace will no longer be priced per agent, but instead based on the number of traces, number of logs, and log retention period that customers use, as detailed on Retrace’s pricing page.

Source: Stackify.

Our Take

This change from an agent-based to a trace- and log-based pricing model is in line with other changes Info-Tech has observed in enterprise IT, both in technology usage and in changing architectures.

As is evident in the rise of cloud services such as AWS Lambda and Elasticsearch, many organizations have become more comfortable with paying according to usage-based models such as queries or API calls rather than other arbitrary elements such as the number of users of an application.

The old agent-based model posed a number of problems – for instance, organizations that run larger numbers of smaller instances in the cloud to use for horizontal scaling end up paying more in licensing than if they were to vertically scale their cloud instances by increasing the CPU and memory of the individual instances.

Some users might be inclined to try to save money on licensing by packing multiple applications onto a single machine – something that was common back in the days of physical servers on premises but was quickly abandoned with the rise of virtualization allowing for improved segmentation of applications and services and simplified administration of the VMs.

In short, a per-agent model means that a customer running the same workloads and using Retrace for the same use cases might pay vastly different fees depending on the customer’s underlying architecture. And this doesn’t even address the added complexities of the agent-based model if the organization is running applications in containers.

Abandoning the agent-based model in favor of a usage-based model makes a good deal of sense for Stackify’s customers. Indeed, in a briefing with Info-Tech, Stackify representatives noted that this change was driven by comments that customers had, as well as by Stackify’s mission to make the solution more widely available by making it more affordable for individual developers.


Want to Know More?

Establish a Program to Enable Effective Performance Monitoring