Sentry MCP

import { describe, it, expect, vi } from "vitest"; import getDoc from "./get-doc.js"; describe("get_doc", () => { it("returns document content", async () => { const result = await getDoc.handler( { path: "/product/rate-limiting.md", }, { constraints: { organizationSlug: null, }, accessToken: "access-token", userId: "1", mcpUrl: "https://mcp.sentry.dev", }, ); expect(result).toMatchInlineSnapshot(` "# Documentation Content **Path**: /product/rate-limiting.md --- # Project Rate Limits and Quotas Rate limiting allows you to control the volume of events that Sentry accepts from your applications. This helps you manage costs and ensures that a sudden spike in errors doesn't consume your entire quota. ## Why Use Rate Limiting? - **Cost Control**: Prevent unexpected charges from error spikes - **Noise Reduction**: Filter out repetitive or low-value events - **Resource Management**: Ensure critical projects have quota available - **Performance**: Reduce load on your Sentry organization ## Types of Rate Limits ### 1. Organization Rate Limits Set a maximum number of events per hour across your entire organization: \`\`\`python # In your organization settings rate_limit = 1000 # events per hour \`\`\` ### 2. Project Rate Limits Configure limits for specific projects: \`\`\`javascript // Project settings { "rateLimit": { "window": 3600, // 1 hour in seconds "limit": 500 // max events } } \`\`\` ### 3. Key-Based Rate Limiting Rate limit by specific attributes: - **By Release**: Limit events from specific releases - **By User**: Prevent single users from consuming quota - **By Transaction**: Control high-volume transactions ## Configuration Examples ### SDK Configuration Configure client-side sampling to reduce events before they're sent: \`\`\`javascript Sentry.init({ dsn: "your-dsn", tracesSampleRate: 0.1, // Sample 10% of transactions beforeSend(event) { // Custom filtering logic if (event.exception?.values?.[0]?.value?.includes("NetworkError")) { return null; // Drop network errors } return event; } }); \`\`\` ### Inbound Filters Use Sentry's inbound filters to drop events server-side: 1. Go to **Project Settings** → **Inbound Filters** 2. Enable filters for: - Legacy browsers - Web crawlers - Specific error messages - IP addresses ### Spike Protection Enable spike protection to automatically limit events during traffic spikes: \`\`\`python # Project settings spike_protection = { "enabled": True, "max_events_per_hour": 10000, "detection_window": 300 # 5 minutes } \`\`\` ## Best Practices 1. **Start Conservative**: Begin with lower limits and increase as needed 2. **Monitor Usage**: Regularly review your quota consumption 3. **Use Sampling**: Implement transaction sampling for high-volume apps 4. **Filter Noise**: Drop known low-value events at the SDK level 5. **Set Alerts**: Configure notifications for quota thresholds ## Rate Limit Headers Sentry returns rate limit information in response headers: \`\`\` X-Sentry-Rate-Limit: 60 X-Sentry-Rate-Limit-Remaining: 42 X-Sentry-Rate-Limit-Reset: 1634567890 \`\`\` ## Quota Management ### Viewing Quota Usage 1. Navigate to **Settings** → **Subscription** 2. View usage by: - Project - Event type - Time period ### On-Demand Budgets Purchase additional events when approaching limits: \`\`\`bash # Via API curl -X POST https://sentry.io/api/0/organizations/{org}/quotas/ \\ -H 'Authorization: Bearer <token>' \\ -d '{"events": 100000}' \`\`\` ## Troubleshooting ### Events Being Dropped? Check: 1. Organization and project rate limits 2. Spike protection status 3. SDK sampling configuration 4. Inbound filter settings ### Rate Limit Errors If you see 429 errors: - Review your rate limit configuration - Implement exponential backoff - Consider event buffering ## Related Documentation - [SDK Configuration Guide](/platforms/javascript/configuration) - [Quotas and Billing](/product/quotas) - [Filtering Events](/product/data-management/filtering) --- ## Using this documentation - This is the raw markdown content from Sentry's documentation - Code examples and configuration snippets can be copied directly - Links in the documentation are relative to https://docs.sentry.io - For more related topics, use \`search_docs()\` to find additional pages " `); }); it("handles invalid path format", async () => { await expect( getDoc.handler( { path: "/product/rate-limiting", // Missing .md extension }, { constraints: { organizationSlug: null, }, accessToken: "access-token", userId: "1", }, ), ).rejects.toThrow( "Invalid documentation path. Path must end with .md extension.", ); }); it("handles API errors", async () => { vi.spyOn(global, "fetch").mockResolvedValueOnce({ ok: false, status: 500, statusText: "Internal Server Error", } as Response); await expect( getDoc.handler( { path: "/product/test.md", }, { constraints: { organizationSlug: null, }, accessToken: "access-token", userId: "1", }, ), ).rejects.toThrow(); }); it("validates domain whitelist", async () => { // Test with absolute URL that would resolve to a different domain await expect( getDoc.handler( { path: "https://malicious.com/test.md", }, { constraints: { organizationSlug: null, }, accessToken: "access-token", userId: "1", }, ), ).rejects.toThrow( "Invalid domain. Documentation can only be fetched from allowed domains: docs.sentry.io, develop.sentry.io", ); }); it("handles timeout errors", async () => { // Mock fetch to simulate a timeout by throwing an AbortError vi.spyOn(global, "fetch").mockImplementationOnce(() => { const error = new Error("The operation was aborted"); error.name = "AbortError"; return Promise.reject(error); }); await expect( getDoc.handler( { path: "/product/test.md", }, { constraints: { organizationSlug: null, }, accessToken: "access-token", userId: "1", }, ), ).rejects.toThrow("Request timeout after 15000ms"); }); });