Reduce token pressure in Daily Observability Report workflow

[Copilot]

✨ Enhancement

What does this improve?

The Daily Observability Report workflow was exhausting its 25M effective-token budget while analyzing weekly AWF firewall and MCP gateway coverage. This change raises the workflow ET ceiling and, more importantly, narrows the prompt so the agent samples and prioritizes runs instead of expanding across the full workflow set.

Why is this valuable?

The workflow keeps its observability objective, but does so with bounded fan-out and clearer stopping conditions. That reduces the chance of repeating token-budget failures while preserving coverage of the highest-signal runs.

Implementation approach:

• Budget control

  • add max-effective-tokens: 40000000 to the workflow frontmatter
  • carry the higher ET limit through the compiled lock file

• Prompt scope reduction

  • remove the initial “list all workflows” step
  • start with one repository-wide logs call instead of workflow-by-workflow expansion
  • reduce the initial fetch size from 100 runs to 30

• Bounded follow-up

  • allow targeted follow-up only when a required run category is missing
  • cap targeted fetches and make them additive to, not exempt from, the overall analysis limit

• Explicit prioritization

  • cap analysis at 20 total runs
  • prioritize failed/cancelled runs, then firewall-enabled runs, then MCP-enabled runs, then recent healthy runs
  • prefer breadth across workflows over repeated sampling from the same workflow

• Prompt clarity

  • document why the workflow uses a higher ET cap
  • clarify that the reduced fetch size is intended to provide a representative recent sample

Example of the new control points:

timeout-minutes: 45
max-effective-tokens: 40000000

{
  "workflow_name": "",
  "count": 30,
  "start_date": "-7d",
  "parse": true
}

[Copilot]

Pull request overview

This PR reduces token pressure in the Daily Observability Report workflow by increasing the effective-token ceiling and narrowing the report prompt to sampled, prioritized workflow-run analysis.

Changes:

• Raises the workflow max effective-token limit from 25M to 40M.
• Replaces broad workflow enumeration with a single capped logs fetch plus limited targeted follow-up.
• Adds explicit run-analysis caps and prioritization guidance.

Show a summary per file

File	Description
.github/workflows/daily-observability-report.md	Updates workflow frontmatter and prompt instructions for bounded observability analysis.
.github/workflows/daily-observability-report.lock.yml	Carries the updated token limit and regenerated compiled workflow metadata/configuration.

Copilot's findings

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 2/2 changed files
• Comments generated: 2