MonitorOps

Use this pattern when you want a scheduled workflow to inspect other agentic workflows using workflow logs and auditing, summarize what happened, and escalate unusual cost or failure patterns.

The agentic-ops repository provides the reference implementation for this approach.

flowchart LR
    schedule([Schedule]) --> analyze[Analyze workflow logs]
    analyze --> report[Report to Discussion]
    analyze --> escalate[Escalate failures to issue]

What this pattern does

This pattern reviews workflow logs across a repository, classifies notable behavior, and publishes a structured report. When it detects repeated failures, abnormal token consumption, or other unhealthy patterns, it can escalate those findings into issues for follow-up.

This pattern is useful for repository-wide monitoring because it creates a durable operational record instead of relying on ad hoc inspection of individual workflow runs.

Typical workflow

1. Run on a schedule to collect recent workflow activity.
2. Analyze logs, costs, and failure signals across runs.
3. Post a summary report to a GitHub Discussion or another durable destination.
4. Open or update issues when the same problem crosses a threshold.

When to use it

Use this pattern when a repository has enough workflow activity that maintainers need a regular summary instead of checking each run manually. It also helps when workflows span multiple teams and failures or waste need to be surfaced in a shared location.

Related Documentation

• BatchOps — Process large volumes in parallel chunks
• Audit Commands — Investigate individual runs and regressions
• OpenTelemetry — Workflow telemetry and spans
• Cache Memory — Persistent state across runs
• Concurrency — Prevent overlapping workflow runs
• Monitoring with Projects — Durable tracking with Projects