Papertrail MCP. Find errors by name. Manage logs with chat commands.

You connect your AI agent right here when you need instant visibility into cloud logs and real-time troubleshooting, all without touching a dashboard. This server lets your client interact with the core functions of log management using pure conversation. You'll use it to search massive event databases, track which systems are sending data, structure your log sources, and control where that data ends up.

To start, when you need to find something specific—whether it’s a bug from last night or an error happening right now—you run search_events. This tool lets your agent search through millions of historical or live log entries using complex query syntax. You can pull back all matching events across defined time ranges, and it handles pagination for you if the results are huge.

If you need to watch logs as they come in, it tails real-time data streams directly into your chat interface.

Need to know what's reporting? Use list_systems. This tool gives you a clean list of every unique system ID that’s currently configured to send logs into Papertrail. It keeps you in the loop on which services are active and sending data, so you never wonder if something important is falling through the cracks.

When your infrastructure gets complex, you gotta organize it. You use list_groups to see every existing organizational log group that contains systems. Then, if a new service pops up or you need to separate staging from production logs, you run create_group. This establishes a clean, logical boundary—a dedicated log group—that keeps related systems and their logs together for easy filtering later on.

Managing where your data goes is just as important. You check what's available by calling list_destinations, which retrieves every configured endpoint for archiving or forwarding data in the account. If you need to send those logs somewhere else—say, a dedicated compliance bucket—you use create_destination. This sets up a new, defined endpoint so your log data gets archived or forwarded exactly where it needs to go.

When you combine these tools, you're talking pure DevOps visibility through natural conversation. You don't gotta touch a single dashboard. Your agent first uses list_systems to verify the sources. Then, if necessary, it runs create_group to organize them into a specific log group. When you’re ready for an answer, it executes search_events, letting you drill down using powerful query syntax against that organized data set.

The entire process—from checking system status to querying historical events and setting up new archival sinks—runs entirely through your AI client's conversation with these tools.