Finding out why their own tool was hard to fit

01 — The challenge

Where it started.

The vendor could not see where engineers were losing time when fitting their APM tool to a customer application, which made the gap between a working install and a difficult one hard to close.

The problem was not the product’s capability but the distance between it and the person configuring it.

02 — What we built

The work itself.

Fit it to real applications. Deployment and testing across applications on different stacks — Java, Python and Node agents — rather than reasoning about it from the documentation.

All the way to code level. The workflow followed from first configuration through to code-level analysis of a highly distributed application, which is where it actually breaks.

Interface changes, and scripts. Specific interface recommendations for reading topology at a glance, plus custom scripts that take steps out of instrumentation.

03 — Results

What the client has now.

A documented account of where configuration fails, per technology, rather than per complaint. Topology and application issues made readable in a few clicks. Instrumentation shortened by scripts rather than by documentation. The vendor’s own team shipped a release carrying the changes.