Rollbar

@rollbar show me occurrence metrics for the last 7 days grouped by environment, and highlight any unusual spikes in production errors

Based on Rollbar's occurrence data for the past week, your production environment logged 1,247 occurrences (up 340% from the prior week), staging had 89 occurrences (stable), and development had 156 occurrences (down 12%). The production spike began 3 days ago and correlates with your v2.4.1 deployment. The top error is 'NullPointerException in PaymentProcessor' accounting for 67% of new production occurrences. This example illustrates how the MCP would surface aggregated metrics; actual counts depend on your project's error volume.

This example uses the Get Occurrences Metrics tool to analyze error patterns over time. Requires read access to the project. The MCP returns raw counts and groupings; the AI synthesizes trends and correlates them with deployment context you provide. Rate limits apply to metric queries on high-traffic projects.

@rollbar check if the mobile-team is assigned to the ios-app project, and if not, assign them with standard access

I checked the team-project assignment using Rollbar's API. The mobile-team is not currently assigned to the ios-app project. However, I cannot perform the assignment action directly — the MCP provides a read-only check tool (Check Team Project Assignment) but does not expose a write endpoint for assigning teams to projects. You'll need to assign the team manually in Rollbar's dashboard under Project Settings > Access, or use Rollbar's REST API directly if you have admin credentials.

This example demonstrates a limitation: the MCP includes verification tools (Check Team Project Assignment, Check User Team Assignment) but lacks corresponding write operations for team management. Useful for auditing access, but you'll need Rollbar's UI or direct API calls for modifications. Auth scope must include team read permissions.

@rollbar pull time-to-resolution metrics for all projects and draft a monthly engineering health report focusing on which teams are resolving issues fastest

I retrieved TTR metrics across your 8 active projects. The backend-services project (owned by platform-team) has a median resolution time of 4.2 hours, the fastest in your account. The legacy-api project (owned by integrations-team) has a median TTR of 31 hours, primarily driven by 12 long-lived issues tagged 'technical-debt'. For your report: platform-team is resolving 89% of issues within 24 hours, while integrations-team resolves 54% in that window. This illustrates how the MCP would return project-level TTR data; actual metrics reflect your team's issue lifecycle.

This example combines the Get TTR Metrics tool with the AI's summarization to generate actionable insights. The MCP returns raw resolution time data per project; the AI interprets it in context of team ownership (which you'd provide or the AI infers from project names). Requires read access across all projects you want to analyze.

When Rollbar MCP speeds up weekly on-call transitions

A 6-person backend team rotates on-call every Monday. The outgoing engineer needs to brief the incoming one on active incidents, resolution trends, and which projects are noisiest. With the Rollbar MCP, the handoff doc writes itself: pull TTR metrics across all projects to flag slow-resolving issues, check occurrence counts grouped by environment to spot staging vs. production patterns, and verify team assignments so the new on-call knows who owns what. This works when your error volume is under 10k occurrences per week—above that, the occurrence metrics tool starts timing out on broad queries. If your handoff is a 15-minute Slack call and you need the data in under 60 seconds, this MCP closes the loop.

When this MCP helps support teams explain outages

A 3-person support team at a SaaS company gets flooded with tickets after a production deploy goes sideways. They need to tell customers which features broke, when errors started spiking, and whether the issue is resolved. The Rollbar MCP lets them query occurrence metrics filtered by deploy timestamp and project, then cross-reference team assignments to confirm which engineering squad fixed it. The Get Project tool surfaces deploy metadata so support can say 'the 2:14 PM deploy to checkout service caused a 300% error spike, rolled back at 2:47 PM.' This scenario falls apart if your Rollbar account has more than 50 projects—List All Projects becomes a pagination nightmare. For small-to-midsize ops teams doing post-mortems in real time, it's a clean win.

When Rollbar MCP turns error data into sprint insights

A 5-person product team runs two-week sprints and wants to track whether new features are introducing regressions. At retro, they pull TTR metrics for the last sprint's projects, compare occurrence counts week-over-week, and check if any team member is bottlenecking resolutions by querying user assignments. The MCP's Get Occurrences Metrics tool groups errors by level (warning vs. critical) so the team can decide if they're shipping too fast. This works when you have fewer than 10 active projects per sprint—beyond that, the manual project ID lookups get tedious because the MCP doesn't batch-query projects by tag or label. If your retros are data-driven and you want error trends in the same workspace as your Linear tickets, this MCP earns its API key.