HyperDX MCP. Debug infrastructure, manage alerts, and check logs from your chat.
Gemini Works with every AI agent you already use
…and any MCP-compatible client
Just plug in your AI agents and start using Vinkius.
HyperDX (Open Source Observability) MCP Server lets your AI agent search logs, manage alerts, and inspect dashboards directly. Connect to your infrastructure to debug issues and track performance regressions using natural conversation.
You can list logs with specific filters, create alert rules, and view metrics from existing dashboards—all without leaving your chat interface.
What your AI agents can do
Create alert
Sets up a new alert rule based on defined system conditions.
Delete alert
Removes a specific alert rule using its unique ID.
Get dashboard
Retrieves the detailed metrics and visualization data for a named dashboard.
You ask the agent for logs (e.g., 'all errors in the auth service last hour'), and it runs the query and returns the structured data.
You tell the agent to 'create an alert if CPU usage hits 90%' or 'delete the old rate limit alert,' and it handles the rule change.
You ask for details on a specific dashboard (e.g., 'show API performance dashboard'), and the agent pulls the relevant metrics and trends.
You specify a time window (like 'last 15 minutes' or 'yesterday') in your prompt, and the agent applies the precise time filters to the query.
Ask AI about this MCP
Supported MCP Clients
Gemini Waiting for input…
HyperDX MCP Server: 7 Tools for Observability
These tools let your AI agent interact with your infrastructure's monitoring data, allowing you to search logs, manage alerts, and view dashboards via natural language prompts.
019e5d24create alert
Sets up a new alert rule based on defined system conditions.
019e5d24delete alert
Removes a specific alert rule using its unique ID.
019e5d24get dashboard
Retrieves the detailed metrics and visualization data for a named dashboard.
019e5d24list alerts
Shows a complete list of all alert rules currently configured in the system.
019e5d24list dashboards
Returns a list of all available monitoring dashboards in the organization.
019e5d24list events
Retrieves structured logs or spans (events) based on specified query filters.
019e5d24list logs
Gets a list of application logs using search criteria like service or error level.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on every call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with HyperDX (Open Source Observability), then connect any of our 4,700+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 4,700+ others, all in one place
- Add new capabilities to your AI anytime you want
- Every connection is secured and compliant automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog every week
What you can do with this MCP connector
HyperDX Observability MCP Server lets your AI agent search logs, manage alerts, and inspect dashboards directly. You can connect it to your infrastructure to debug issues and track performance regressions using natural conversation. You'll use it to list logs with specific filters, create alert rules, and view metrics from existing dashboards—all without leaving your chat interface.
Querying Logs and Events
When you ask the agent for logs (like, 'all errors in the auth service last hour'), it runs the query and returns the structured data. You can list events using list_events based on query filters. You'll get a list of application logs using list_logs when you specify search criteria like service or error level.
You can also list all application logs using list_logs if you don't specify criteria.
Inspecting Dashboard Metrics
If you need details on a specific dashboard (say, 'show API performance dashboard'), the agent pulls the relevant metrics and trends for you using get_dashboard. You can first get a list of all available dashboards in the organization with list_dashboards. You'll get the necessary data to visualize metrics and trends by calling get_dashboard.
Managing Alert Rules
To keep tabs on system performance, you can list all configured alert rules using list_alerts. If you spot a performance problem, you'll set up a new alert rule based on defined system conditions using create_alert. If that rule becomes outdated, you can remove it using delete_alert with its unique ID.
How HyperDX MCP Works
- 1 First, subscribe to the server and enter your HyperDX API Key.
- 2 Next, prompt your AI agent with a natural language request (e.g., 'List all error logs for the user service from 15 minutes ago').
- 3 Finally, the agent executes the necessary HyperDX tool call, retrieves the data, and formats the results for you to read in the chat.
The bottom line is you never have to leave your chat window to correlate logs, events, and alerts.
Who Is HyperDX MCP For?
The DevOps engineer who's tired of clicking through dashboards at 2 am. The Site Reliability Engineer who needs to manage alert thresholds on the fly. Developers who want to grab logs for a specific trace ID without leaving their coding flow. If you live in a terminal or chat interface, this is for you.
Checks error rates and system logs instantly without switching to a browser or dedicated observability UI.
Investigates incidents and manages alert thresholds directly from the chat or terminal interface.
Fetches logs for specific services or trace IDs while staying in the flow of coding or debugging.
What Changes When You Connect
- Cross-reference everything without clicking. Use the agent to run
list_logsandlist_eventsin sequence, correlating a specific error log entry with the related system event that triggered an alert. - Manage system health rules on the fly. Use
list_alertsto see what's configured, andcreate_alertto set a new threshold (e.g., 'if errors > 50 in 5m')—all from your chat. - See the big picture quickly. Run
list_dashboardsto find all available views, then useget_dashboardto pull specific metrics and trends without opening the web UI. - Filter data precisely. Instead of just listing everything, use
list_logswith filters likelevel:error service:authto narrow down the search and focus only on what matters. - Work with time ranges. You can query data using relative time periods (like 'last 24 hours') or specific ISO 8601 timestamps, ensuring your troubleshooting is precise.
- Stay in flow. The agent runs the query, returns the data, and you keep your focus on the task at hand—whether that's coding or writing a post-mortem.
Real-World Use Cases
Diagnosing a Production Outage
The site is down. Instead of clicking into the dashboard, the engineer asks the agent: 'Show me all error logs and related events for the payment service from the last hour.' The agent runs list_logs and list_events, correlating the data and showing the root error pattern that triggered the high-severity alert.
Setting up New Monitoring
The team needs better visibility on a new microservice. The developer tells the agent: 'Create an alert if the latency for the user profile service exceeds 500ms.' The agent uses create_alert to build the rule and sends confirmation, keeping the ops team fully informed.
Auditing System Behavior
A PM asks why user signups dropped last week. The SRE asks the agent to use list_events to pull all structured activity data for the last seven days. The agent retrieves the events, allowing the team to pinpoint when the drop began and why.
Investigating a Sporadic Bug
A developer is debugging a rare bug. They prompt the agent: 'Get me all logs for trace ID XYZ from the last 30 minutes.' The agent runs list_logs, filtering by the specific trace ID, allowing the developer to see the exact sequence of calls that failed.
The Tradeoffs
Manual Data Stitching
The developer sees a spike in the dashboard. They open the log view, then open the event view, then open the alert list, and manually try to match timestamps across three different tabs.
→
Instead, ask the agent to run a correlated query. The agent handles the logic behind the scenes, pulling and aligning the data from list_logs and list_events into one unified response.
Forgetting Time Filters
Running list_logs without specifying a time range returns thousands of entries, making the results useless noise. The engineer has to scroll for hours to find the relevant incident.
→ Always specify the time. Use the agent's ability to filter data by relative time ranges (e.g., 'last 1h') or specific timestamps to keep the search focused.
Over-relying on UI defaults
Assuming the dashboard shows the full picture. The user sees a metric drop but doesn't know why. They rely on the dashboard's default view, which only shows the aggregated number.
→
Use list_dashboards to find the right dashboard, then use get_dashboard to fetch the raw data details. You can then ask the agent to interpret why the metric changed based on the retrieved data.
When It Fits, When It Doesn't
Use this if your primary job is incident response, debugging, or maintaining system health. You need to correlate data streams (logs, events, alerts) and you hate switching browser tabs. You're looking for the root cause, not just a number.
Don't use this if you just need to view static, historical reports that don't require immediate correlation. If you just need a simple list of resource names, a basic query tool is enough. This server is for deep, dynamic investigation—it's about tracing causality, not viewing status reports.
When in doubt, ask your agent to combine tools. Example: 'What alerts fired (using list_alerts) related to service failures (using list_logs) in the last 24 hours?'
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by HyperDX. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on every call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
Works with Claude, ChatGPT, Cursor, and more
The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.
This server provides 7 capabilities that interface natively with Claude, ChatGPT, Cursor, and any MCP client. No middleware. No custom integration required.
Available Capabilities
Debugging system failures shouldn't require opening five different tabs.
Today, finding the root cause is a mess. You see an alert fires. You open the dashboard to check the metric. Then you jump to the logs to see the stack trace. Next, you go to the event stream to see if a background job failed. You copy three timestamps and paste them into a spreadsheet just to try and align the data. It's slow, it's painful, and you lose context.
With this MCP server, you just tell your agent what you need. You ask, 'Why did the error rate spike at 2:15 PM?' The agent runs the tools—it checks the alerts, pulls the logs, and cross-references the events—and gives you a single, actionable answer. The whole process stays in your chat.
HyperDX MCP Server: Manage Alerts and Dashboards from Chat
Before this, managing alerts meant logging into the dedicated monitoring UI, finding the correct rule, and manually editing the threshold. If you needed to check what alerts existed, you had to navigate through menus to see the list. It was a multi-step, browser-dependent chore.
Now, you just ask the agent to 'List all alerts' or 'Create an alert if latency > 500ms.' The agent executes the `list_alerts` or `create_alert` tool and confirms the change instantly. The entire operational cycle is contained within your chat interface.
Common Questions About HyperDX MCP
How do I use the `list_logs` tool to find specific errors? +
You use list_logs by providing specific filters in your prompt. For example, ask for 'logs with level:error for the auth service.' The tool handles the query structure, pulling only the relevant records.
Can I create an alert using the `create_alert` tool? +
Yes. You simply ask the agent to create the alert rule, specifying the condition (e.g., 'errors > 50 in 5 minutes') and the scope. The agent calls create_alert for you.
What is the difference between `list_events` and `list_logs`? +
Logs are application-level records, usually containing stack traces. Events are structured logs or spans that track the flow of an action across services. Use list_logs for stack traces, and list_events for process flow.
How do I view metrics using the `get_dashboard` tool? +
You ask for the dashboard by name. The agent uses get_dashboard to pull the specific metrics and trends, presenting the visualization data right in the chat window.
How do I check if I can delete an alert using the `delete_alert` tool? +
Yes, you use delete_alert to remove an alert rule. You simply provide the specific alert ID when calling the tool.
What time formats can I use when querying data with `list_logs`? +
You can use relative time ranges (like '15m' or '1h') or precise ISO 8601 timestamps. This gives you precise control over your data window.
Can the `list_dashboards` tool retrieve only specific types of dashboards? +
The list_dashboards tool retrieves all dashboards configured in your organization. You then use get_dashboard to pull the details for a specific one.
If my log search fails, how can I adjust my query for `list_logs`? +
You adjust your query parameters directly in the tool call. For example, you can narrow the scope by adding a service name or a specific log level.
Can I search for specific errors in my logs using this server? +
Yes! Use the list_logs tool with a query like level:error. You can also specify a time range using the from parameter (e.g., '1h' or '24h') to narrow down the results.
How do I set up a new alert for a specific service? +
You can use the create_alert tool. You'll need to provide a name, the search query (e.g., service:auth level:error), a threshold value, the type of alert (like 'count'), and the evaluation interval (e.g., '5m').
Is it possible to delete an alert rule if it's no longer needed? +
Yes, simply use the delete_alert tool and provide the unique ID of the alert rule you wish to remove.
Use it with your favorite AI tools
Connect this server to Cursor, Claude, VS Code, and more.
More in this category
Incident.io
Manage incidents, roles, and on-call schedules via Incident.io API.
Fly.io Alternative
Manage edge compute infrastructure via Fly.io — list apps, control machines (VMs), manage volumes and certificates from any AI agent.
NS1 (IBM NS1 Premium DNS API)
Manage DNS zones, records, and traffic steering via IBM NS1. Automate infrastructure directly from your AI agent.
You might also like
Retell AI
Build human-like AI voice agents that handle phone calls, answer questions, and complete tasks through natural spoken conversation.
Sunrise-Sunset
Get precise sunrise, sunset, and twilight times for any location on Earth using latitude and longitude coordinates.
Mapbox
Geocode addresses, get directions, calculate distances and generate maps with Mapbox's location platform.