Where Ketl is going

Deeper GitHub deploy correlation

Linking the specific commit, author, and changed files to the alert window, so deploy evidence in a diagnosis is precise rather than approximate.

Slack delivery GA

Posting the ranked cause and proposed fix into the incident channel the moment a diagnosis completes. Includes a follow-up thread for additional evidence.

More log sources

Elastic as a first-class log source with the same ingest quality as Datadog and Loki, added alongside existing connectors.

Per-source read window controls

Let teams set the look-back window per source so a noisy log store does not crowd out a sparse but important one.

Opt-in auto-remediation

Per-action, human-approved remediation: a rollback command runs only after an engineer confirms it. Starting with the safest, most reversible actions.

Opsgenie and CloudWatch alarm triggers

Two of the most-requested alert sources, added as trigger connectors alongside PagerDuty.

Incident timelines

A visual timeline of events across all signals for a given alert window, so you can orient in the sequence before reading the ranked cause.

Dedicated API tokens for Team

Per-team, per-rotation API tokens with scoped permissions, rather than a single account credential.

Self-host GA

Self-hosted deployment as a first-class, fully supported option for teams who need telemetry to stay entirely on-premises.

SSO and SCIM GA for Scale

Enterprise identity management: single sign-on, SCIM-based provisioning, and group-based access control, generally available.

Routing to customer-hosted open models

Teams on Scale can already route to open models they host. Making this configuration stable, documented, and generally available.

Learned per-service priors

Ketl learning which failure modes are common for each of your services, weighting evidence and confidence to match the history of that service.