What's the difference between a proven cause and a lead in an RCA report?

A proven cause has a located change, a diff, a prompt edit, a retrieval-code change, lining up with the trace evidence and is marked high confidence and carded on the case page; a lead is unconfirmed to that bar, still backed by trace evidence but with no located change, marked medium or low, and shown only when the report has no proven cause for that finding.

Every RCA report now writes the same shape, on a metric-movement case or a frustration or groundedness one: a one-sentence summary, a verdict, and causes carrying a title, what changed, how that caused the shift, and what to do next, each cited to the traces or sessions it draws from, the same standard any regression investigation is expected to meet. That’s new as of a September 2026 update; reports before it wrote a flat, unranked list with no confidence tier at all.

A report with no proven cause and no surviving lead still reads no_cause_found, same as before. Leads are new visibility into what the investigation considered and couldn’t confirm, not a lowered bar for what counts as a named cause.

sources

keep reading

More on this.

Two ways to run Tessary.

Tessary is an open-source agent reliability platform. Cloud and self-hosted run the same workflow on the OpenTelemetry traces your agent already emits.

Tessary Cloud

We host it for you. Send your first trace with nothing to deploy and no model key.

what's includedper organization
traces
10,000 per calendar month
stored trace data
1 GB
retention
30 days
model credit
$10, one-time, for triage and root-cause analysis
credit card
not required

Self-hosted Tessary

Run the open-source code on your own infrastructure with one command. Add your own model key for triage and root-cause analysis.

Self-host Tessary for me by following https://github.com/tessaryai/tessary/blob/main/setup.md

docker compose -f oci://docker.io/tessaryai/tessary:compose up -d -y