Probes that see what
your users see
Synthetic monitoring, browser automation, and end-to-end testing — one platform, native to the Koder Observability stack.
Synthetic monitoring, browser automation, and end-to-end testing — one platform, native to the Koder Observability stack.
Probes simulate real users — continuously in production, on every PR, and on demand for load tests.
Run probes on a schedule, from any region, and alert when production starts behaving differently from how your users expect.
Same probe, run on every PR via koder-flow CI. Failures post a status check with a link to the run waterfall.
Spin up N concurrent virtual users from the same script. Get P50/P95/P99 latency without writing a second test.
Every run produces a narrative log: a step-by-step prose description, plus screenshots, video, HAR, console, DOM, and Web Vitals.
Chromium today, Firefox and WebKit next. Run the same probe on all three and compare results side-by-side.
Write declarative YAML, port your Playwright TS scripts, or record a probe by browsing. All compile to the same IR.
From script to alert in three steps.
Author it in YAML, TypeScript, or just browse with the recorder turned on.
Run on cron in production, on every PR in koder-flow, or on demand from the CLI.
Waterfall, screenshots, video, HAR, console, narrative log — all cross-linked into koder-trace-apm and koder-log.
Probe runs become first-class citizens in every observability service you already use.
Console and narrative logs streamed via the Loki push API.
Every step emits OTLP spans, correlated with backend traces.
Alerts on failure, regression, and degradation.
Prometheus metrics about queue depth, success rate, and worker health.
Delegate complex checks that need a real browser.
OIDC login, multi-tenancy, and RBAC out of the box.