BugTape docs

BugTape and Datadog

Where the agent fix loop fits next to APM.

BugTape and Datadog do different jobs, and many teams run both.

Datadog answers "how is the system doing": APM traces, log search, infrastructure metrics and dashboards across the whole estate.

BugTape answers "what broke for this user, how often, and what is the fix": it records what happened before each failure, groups every report of the same failure into one issue, shows who it reached, and hands the evidence to a coding agent over MCP.

Use BugTape for

• Bugs your users hit, with the page, the failed request, the console and the stack attached.
• One issue per failure, with a count and the users and releases it reached, instead of a pile of tickets.
• Letting a coding agent read a ranked issue (list_recommended_critical → get_repro_context) and open a pull request.
• Server jobs (Python, Node.js) where the value is grouping, who was affected and a repro packet for the agent.

Keep Datadog for

• APM traces, log analytics and infrastructure metrics across your services.

Together

When a report carries a Datadog trace or RUM session id, the BugTape issue links straight to it, so the fix starts with the user's evidence and can drop into the trace in one click. See Use BugTape with Datadog.

When BugTape is not the tool

Warehouse or BI analysis, retention of event evidence beyond 30 days, fatal native crash capture, or screens with client data where masking has not yet been proven with the privacy checklist. For those screens, start with server-side capture only.