A Redis latency bug survived a year of fixes at Shaped.
Then Kathleen Xue, a senior engineer there, ran the traces through a coding agent with the SigNoz MCP server.
Her take 👇
New customer story: Kernel x SigNoz.
We spoke with Hiro Tamada about how Kernel uses SigNoz MCP in production for AI agent infrastructure: post-launch checks, latency investigations, and Slack-based analysis.
One highlight: browser acquisition latency went from 140ms to 30ms.
Five alerts do not always mean five incidents.
With SigNoz MCP, your coding agent can help correlate alerts, identify connected services, and point responders toward the signal that likely matters first.
Find the full use case doc below 👇
If you are starting with OpenTelemetry, this might be the only video you need.
It covers how it started, what problems it solved, the architecture, a collector deep dive all tied together with a live demo. Everything straight from the official docs and covered within 15 minutes.
Good alerts start as plain-language intent:
"Tell me when checkout errors stay high after a deploy."
With SigNoz MCP, your coding agent can help turn that into an alert workflow you can review before enabling.
Find the full use case doc below 👇
