This website uses cookies
Read our Privacy policy and Terms of use for more information.
Approvals are working fine. It is the other end of the process that nobody owns.
Oct 9, 2026
•
4 min read
An agent's run mode, a CPU bill, a week of redeploys. September was about costs nobody put their name to.
Oct 2, 2026
5 min read
Your agent estate is your org chart with a bill attached.
Sep 27, 2026
6 min read
Issue 106. Whole-system profiling, and the cycles on your bill your team never shipped
Sep 25, 2026
OpenTelemetry's fourth signal shows which code is spending the cycles you already pay for, including code that is not yours.
Sep 22, 2026
Issue 105. What under-instrumenting actually costs you, and how to price yours.
Sep 18, 2026
Autonomy shipped as a setting. Settings have owners.
Sep 14, 2026
Issue 104. Agent-driven journeys and the synthetic checks that assume a human.
Sep 11, 2026
Google and Microsoft shipped a way for websites to hand agents a list of tools. Your scripted checks assume a user with eyes.
Sep 8, 2026
7 min read
The blocker is never the tooling. It is the four arguments nobody schedules.
Sep 4, 2026
Everyone is buying the autonomous layer. August was about the unglamorous thing it stands on.
Sep 1, 2026
The tooling has been fine for years. The blocker is the agreement: what counts as a journey, who owns it, and whether the estate data can be trusted.
Aug 30, 2026
Dynatrace built a mode with no human in the loop. Read its own safeguards and the gate is still there. It moved.
Aug 29, 2026
One tells you when something you expected breaks. The other lets you ask why, and what else, when something you did not expect breaks.
2 min read
Everyone prices the licence. Nobody prices the query log.
Aug 28, 2026
3 min read
The AIOps and service-management platform for the enterprise that never got to greenfield, and the honest question to ask before you consolidate onto it.
Aug 27, 2026
Everyone demos the reasoning. Nobody demos what the agent does to your data layer, or what happens when it hits the rate limit mid-incident.
Aug 23, 2026
Two observability acquisitions in six days. Here is what that means for who owns your telemetry.
Aug 21, 2026
The demo shows you the tower. It never shows you the year underground.
Aug 9, 2026
The five levels are not a scorecard to feel good about. They are a map of what to fix next.
Observability you can act on, whether you build it or lead it, plus the few signals worth your time.
Aug 2, 2026
Continuous profiling is becoming observability's fourth signal. What it is, what it does well, and the honest cost of always-on.
Jul 30, 2026
A governance model for agent actions that fits in your head: decide by blast radius, not confidence.
Jul 26, 2026
Not 'is the model up', but 'why did the agent do that'. Seeing an autonomous agent's model calls, tool calls and reasoning as one trace.
Jul 23, 2026
Two very different kinds of AI in operations. One is fluent, one reasons about cause. Which you trust for a root cause decides whether you fix the incident or a coincidence.
Jul 22, 2026
