docs(TSP-1305): add Monitor alarms to Evals documentation

[claude]

Summary

• Adds a new Setting up alarms subsection under the Monitor section of build/agents/build-your-agent/evals.mdx
• Documents per-check alarms (TSP-1303): set thresholds on individual Check pass rates with M-of-N bucketed breach logic
• Documents overall Monitor alarms (TSP-1305): set thresholds on the aggregate pass rate across all Checks in a Monitor
• Covers notification channels (email, Slack, Microsoft Teams) with links to integration docs
• Includes a step-by-step UI configuration guide

Linear issue

https://linear.app/relevance/issue/TSP-1305/

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated (UTC)
relevanceai	🟢 Ready	View Preview	Jun 17, 2026, 3:45 AM

💡 Tip: Enable Workflows to automatically generate PRs for you.

[linear]

TSP-1305

[github-actions]

🎯 Vibe check

Reviewed: 1 file (1 with issues, 0 clean)

Scores

	Dimension	Score	What's holding it back
🟢	Consistency	9/10	evals.mdx:8 — bold label **Rollout Status** inside a callout violates CLAUDE.md rule ("no bold labels inside" callouts). No banned words, no British spellings, capitalization is otherwise clean.
🟡	Technical clarity	8/10	evals.mdx:312 — "M-of-N bucketed breach logic" is unexplained jargon dropped without definition; the plain-language sentence that follows it is the useful part and should stand alone.
🟢	Non-technical clarity	9/10	Overview and section intros are well-written. Jargon is introduced gradually. Minor: "M-of-N bucketed breach logic" (line 312) would also confuse non-technical readers.
🟡	Structure	7/10	Best practices section uses <CardGroup> where CLAUDE.md says best practices should be bullets. Five sequential step-by-step procedures use plain numbered lists where <Steps> would give visual progress indicators.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

✨ Overall vibe: Solid, comprehensive documentation for a complex feature — well-organized, specific UI references throughout, and the FAQs section handles edge cases cleanly. The main friction points are structural: several sequential procedures that would benefit from <Steps>, and a best practices <CardGroup> that CLAUDE.md explicitly says should be a bullet list instead.

🔧 Issues (2)

• build/agents/build-your-agent/evals.mdx:8 — **Rollout Status** is a bold label inside a <Info> callout. CLAUDE.md says callouts must be "a single short paragraph — no bullet lists, no multi-line content, no bold labels inside." Remove the bold label and fold the text into a plain sentence: e.g., Evals is rolling out progressively, starting with Enterprise customers. If you don't see this feature, reach out to your account manager.
• build/agents/build-your-agent/evals.mdx:312 — "The alarm uses M-of-N bucketed breach logic" introduces undefined jargon. The sentence that follows already explains the behavior clearly. Drop the jargon label: The alarm evaluates the Check's pass rate over the configured time window each time a new result arrives for that Check.

🧩 Component suggestions (2)

• build/agents/build-your-agent/evals.mdx:375–394 — Best practices are in a <CardGroup cols={2}> with 6 cards. CLAUDE.md explicitly says <CardGroup> is not appropriate for "best practices that read naturally as flowing bullets." Convert to a <CardGroup> → bulleted list. Each best practice's title and description paragraph can become a **Start simple** — Begin with a few core scenarios... style bullet, or simply drop the bold titles and write them as clean prose bullets.

• build/agents/build-your-agent/evals.mdx:101–106, 120–151, 214–216, 266–271, 340–347 — Five sequential step-by-step procedures use plain numbered lists. CLAUDE.md says to use <Steps> for sequential procedures. Wrapping each in <Steps> adds visual progress indicators and makes multi-step flows easier to follow. Example for the shortest one (running evaluations, lines 214–216):

  <Steps>
    <Step title="Name the run">Enter a name or accept the default timestamp.</Step>
    <Step title="Add global checks">Select any additional Checks under **Additional global checks**.</Step>
    <Step title="Run">Click **Run** to begin.</Step>
  </Steps>

🏗️ Page structure (1)

• build/agents/build-your-agent/evals.mdx — No closing CTA. This is a concept + reference hybrid for a major new feature; readers who finish the page have learned what Evals does but have no clear next step. Suggest a short ## What's next? section pointing to:
  • /build/agents/build-your-agent/alerts — complement to Monitor alarms, already referenced inline
  • The Agent publishing flow (wherever the Publish/deploy docs live) — the Publish section sets up a natural bridge

⚠️ Contradictions (0)

🔋 Credit usage

Item	Count
Files reviewed	1
Context pages read	2
Total lines processed	~553

Files read: build/agents/build-your-agent/evals.mdx (442 lines), build/agents/build-your-agent/alerts.mdx (111 lines)