diff --git a/workshop/22-error-handling-and-resilience.md b/workshop/22-error-handling-and-resilience.md index 49ef9901..18109ed2 100644 --- a/workshop/22-error-handling-and-resilience.md +++ b/workshop/22-error-handling-and-resilience.md @@ -28,6 +28,14 @@ Agentic workflows can fail for several reasons: Recognising these patterns helps you write instructions that stay on track. +The diagram below shows how these failure modes map to the three mitigations covered in this step. + + + + + Four failure modes — prompt drift, timeout, tool error, and empty data — each mapped to one of three mitigations: defensive brief, timeout-minutes, and fallback safe-output, which together produce a reliably running workflow + + ### Apply all three changes with the skill In the GitHub Copilot **Chat** or **Agents** tab, paste: diff --git a/workshop/images/22-resilience-techniques-dark.svg b/workshop/images/22-resilience-techniques-dark.svg new file mode 100644 index 00000000..0de60df2 --- /dev/null +++ b/workshop/images/22-resilience-techniques-dark.svg @@ -0,0 +1,133 @@ + + + + + + + + + + + + + + + + + + + + Resilience Techniques for Agentic Workflows + + + + + FAILURE MODES + + + MITIGATIONS + + + OUTCOME + + + + + + + Prompt drift + Ambiguous brief → unexpected path + ⚠ + + + + Timeout + Agent reasoning exceeds time limit + ⚠ + + + + Tool error + API rate-limit or network failure + ⚠ + + + + Empty data + No issues or PRs to summarise + ⚠ + + + + + + + + + + + + Defensive task brief + Explicit fallback instruction when data is missing + 🛡 + + + + timeout-minutes + Cap run duration in workflow frontmatter + ⏱ + + + + Fallback safe-output + Always call output — even for quiet runs + 📤 + + + + + + + + + ✅ + Workflow runs + reliably + unattended, week after week + + + + Apply a defensive brief, a timeout, and a fallback safe-output to keep your workflow useful even when things go wrong. + + diff --git a/workshop/images/22-resilience-techniques-light.svg b/workshop/images/22-resilience-techniques-light.svg new file mode 100644 index 00000000..9d756a28 --- /dev/null +++ b/workshop/images/22-resilience-techniques-light.svg @@ -0,0 +1,145 @@ + + + + + + + + + + + + + + + + + + + + Resilience Techniques for Agentic Workflows + + + + + FAILURE MODES + + + MITIGATIONS + + + OUTCOME + + + + + + + Prompt drift + Ambiguous brief → unexpected path + ⚠ + + + + Timeout + Agent reasoning exceeds time limit + ⚠ + + + + Tool error + API rate-limit or network failure + ⚠ + + + + Empty data + No issues or PRs to summarise + ⚠ + + + + + + + + + + + + + + + + + + + Defensive task brief + Explicit fallback instruction when data is missing + 🛡 + + + + timeout-minutes + Cap run duration in workflow frontmatter + ⏱ + + + + Fallback safe-output + Always call output — even for quiet runs + 📤 + + + + + + + + + + + + + + ✅ + Workflow runs + reliably + unattended, week after week + + + + Apply a defensive brief, a timeout, and a fallback safe-output to keep your workflow useful even when things go wrong. + +