Spacelift announced Run Observability, built-in run analysis for OpenTofu that shows platform teams why an OpenTofu run is slow or failing and what to fix. Platform teams can tell within seconds whether a slow run is due to an outage, a provider, or their own code, and catch regressions before app teams start asking why deployments are slow.
Run Observability breaks down OpenTofu runs by phase, time used by resource and provider, and surfaces the real error and flags slow runs without having to re-run it with debug logging. OpenTofu provides the raw trace data. Since version 1.10, OpenTofu can emit OpenTelemetry traces of each run and export them to an OpenTelemetry Protocol (OTLP) backend. Run Observability adds analysis by checking each run’s trace against a set of rules and the stack’s run history, turning the results into findings with actionable recommendations.
“Visibility into infrastructure deployment runs has been a major blind spot. As the third-party technologies around a platform team’s own infrastructure struggle to scale, teams have little insight into where or why things break, and the OpenTofu community has been asking for a solution,” said Jonah Kowall, senior vice president of product and design at Spacelift. “As the leading sponsor of OpenTofu, we helped build tracing into the project to unlock this visibility. Now we’re bringing that capability to Spacelift customers through Run Observability, with no changes required on their end, giving them deeper insight into and more control over their deployments.”
When a run drags or fails, platform teams can’t tell whether OpenTofu, a provider, a third-party API or their own code is responsible. Run logs report a total run time with no breakdown across init, plan and apply, and finding the real error means scrolling through output or re-running with debug logging. Most platforms stop at logs and a total run time. Run Observability shows which phase, resource or provider caused the problem. If GitHub goes down mid-run, for example, the phase breakdown shows initialization running far past normal and names the provider download that stalled, so the team can stop searching its own code.
The volume of deployment runs from AI are also increasing the demand for observability. In Spacelift’s Infrastructure Automation Report 2026, 76% of respondents said they were willing to apply AI-generated Terraform to production with little to no scrutiny, and 93% of organizations reported at least one AI-caused infrastructure incident.
“AI coding tools drive more infrastructure changes and more runs, which pushes platform teams to understand why runs fail and fix them faster,” said Pawel Hytry, CEO of Spacelift. “Run Observability puts the answer on the run itself. When an engineer can see in seconds that a provider download stalled the run, the team stops losing afternoons to run forensics and gets back to building the platform.”
Run Observability is generally available today.
FAQ
Do OpenTofu users need Spacelift to get these traces?
No. OpenTelemetry tracing is part of open source OpenTofu, and any OpenTofu user can send traces to an OTLP-compatible backend such as Jaeger or Grafana Tempo. For Spacelift customers, Run Observability collects and analyzes those traces automatically on OpenTofu native stacks, with no backend or collector to set up.
Does Spacelift’s Run Observability work on Terraform?
No. This new level of visibility and metrics are available exclusively in Spacelift OpenTofu stacks.
Are the findings generated by AI?
No. OpenTofu supplies the raw trace data, and Spacelift supplies the analysis. Each finding comes from a fixed set of rules applied to the run’s trace and the stack’s run history, presented as a natural language explanation with a suggested fix.
