From blind clusters to full visibility -
taking over a platform its builders left behind

Our client acquired a SaaS product in the EdTech space when a much larger company decided it was done with it. More than 50 people had been building the product, which ran across Kubernetes clusters.

With the acquisition, everyone who knew how the system worked was gone. There was barely any monitoring in place: when something broke, engineers pulled logs out of individual pods by hand and hoped the answer was in there. The new owners needed someone to take over the platform - and to make it visible again.

First, as always, an audit: how the system was put together, how to deploy it properly, and what the clusters needed to stay healthy.

Then we stopped flying blind. We set up a full monitoring suite on the LGTM stack - Loki, Grafana, Tempo, and Mimir - covering metrics, logs, and traces. Along the way we found security holes and closed them to protect the system from abuse. We've been running the clusters ever since.

The clusters, nodes, and pods are now fully visible, with metrics, logs, and traces in one place. When something looks off, the answer is on a dashboard - not buried in a pod somewhere. For the owners, the platform finally feels stable and secure - and they're not running it alone: we maintain the clusters for them, day to day.