Skip to content

HOWTO: Interpret Structured Probe Output

Purpose: Read and evaluate structured probe output from HybridOps verification steps to confirm environment health or identify degradation.

Difficulty: Intermediate

Track: Verification-Driven Operations

Overview

Probes are the verification primitives that HybridOps uses to confirm environment state. Unlike raw CLI output, probe results are typed, structured, and threshold-bound: making them machine-readable and audit-suitable. This HOWTO teaches you to navigate probe output as a first-class records, not just as a pass/fail indicator.

1. Where Probe Output Lives

  • Probe records paths under <runtime-root>/logs/.
  • Probe output in the run record run-record output.
  • Locating probes from a specific step.

2. Probe Output Schema

  • Signal types: latency, reachability, count, diff, boolean.
  • Value, unit, threshold, and evaluation fields.
  • Timing: collected_at, duration_ms.

3. Evaluating Threshold Results

  • Pass, warn, and fail thresholds.
  • Absolute vs relative thresholds.
  • Reading the evaluation message.

4. Correlating Across Steps

  • Before/after probe pairs and how to compare them.
  • Cross-step correlation using the shared run ID.
  • Multi-target probe aggregation.

5. Distinguishing Real Failures from Transient Noise

  • Retry behaviour and its effect on probe output.
  • Single-sample vs windowed probes.
  • When to escalate vs re-probe.

References

License: MIT-0 for code, CC-BY-4.0 for documentation unless otherwise stated.