Office hours: is FinOps just observability

Less toolkit walkthrough this time, more a conversation about what FinOps actually is as a discipline, sparked by a question about ingesting Azure AI Foundry cost data.

FinOps as observability

While debating where AI cost data should live (the toolkit’s own store, Log Analytics, or a cross-cluster query), the conversation turned to a bigger claim from Michael Flanakin:

FinOps is really fundamentally an observability story. It’s just observability about cost. There’s no reason cost management couldn’t be merged into monitor.

It’s a reframing worth sitting with: if cost is just another signal alongside performance and reliability, the tools and instincts that already exist for observability should transfer more directly than most FinOps tooling currently assumes.

What AWS gets right that Azure doesn’t yet

One attendee, who’s worked with both clouds, didn’t hold back on the comparison:

I can tell you from AWS experience, getting the billing telemetry was a lot easier. It was night and day different.

Azure’s path to the same dashboards typically runs through Azure Data Explorer, Data Factory, and cost exports, then hand-written KQL, exactly the gap the FinOps toolkit exists to close. It’s a useful reality check on how far the tooling still has to go before “just look at your cost dashboard” is as easy on Azure as it apparently is elsewhere.

Two smaller items worth watching: a skillset integrating the Azure SRE agent with quota and capacity data for automated overspend and quota-pressure alerts, and a fix for savings-plan cost export, a current gap in Cost Management’s native exports.

If you’ve got an opinion on whether FinOps is “just” observability, office hours is a good place to argue about it. Every other Wednesday, open to anyone.