Connectors / Observability
Observability MCP connectors for Claude, ChatGPT, Cursor and Microsoft Copilot
Elaichi connects your team’s AI to 10+ Observability apps. Each person’s AI can do only what they’re allowed to, and every action it takes in your apps is logged.
- Free 14-day trial
- No credit card
- Free plan for one person
Observability connectors in the catalog
Connect one, share it with your team, and each person’s AI can do only what they’re allowed to. Nobody has to hand a login around.
12 connectors
-
Axiom Hosted sign-in 31 tools
-
Azure Resource Manager Hosted sign-in 521 tools
-
Datadog Hosted sign-in 352 tools
-
Grafana Cloud Hosted sign-in 62 tools
-
Honeybadger Hosted sign-in 39 tools
-
Honeycomb Hosted sign-in 34 tools
-
LogRocket Hosted sign-in 8 tools
-
Monte Carlo Hosted sign-in
-
New Relic API key 54 tools
-
Pydantic Logfire Hosted sign-in
-
Splunk API key 112 tools
-
Vantage API key 226 tools
No Observability connector by that name
It may sit under another category in the full catalog. Any documented API can become a connector.
What teams do with Observability connectors
Ask in plain English. Your AI does the work in the Observability apps your team already uses.
Engineering
-
Trace a slow request to its cause
Find the spans behind a slow endpoint, compare them with the metric for that service and see where the time goes.
-
Catch a dashboard up after a deploy
Read the queries behind each panel, then update the dashboard when a metric name or label changed in the release, instead of editing panels one at a time.
-
Map every entity in a workload
List the services, hosts, and other entities grouped under a New Relic workload, along with their tags, to see what a team actually owns.
Support
-
Check whether a customer hit a known fault
When a ticket comes in, ask whether that user appears among the affected users on any open Honeybadger fault, and get the fault's status and the last comment on it.
-
Confirm a customer's reported error
Ask for recent spans from that customer's requests and get a plain answer you can paste into the ticket.
-
Confirm an outage before replying to a ticket
Ask whether requests for a given account failed in the last hour, then reply to the customer with what actually happened.
Platform
-
Open an outbound port for a new destination
Add the outbound port Splunk needs to reach a new forwarding target, list what is already open, and remove ports nobody uses any more.
-
Build a dashboard from a plain question
Describe the dashboard you want, say p95 latency per endpoint for the payments service, and have it created in Pydantic Logfire. Adjust the panels by asking rather than by clicking through settings.
-
Review monitors before they go live
Validate a new monitor against Datadog, compare it with the monitor templates your team already uses and check which monitor group it belongs in.
Security
-
Trace who made a change and when
Pull activity log entries for a resource group filtered to administrative events and see which account changed what, ready to paste into an incident write-up.
-
Review who can see which costs
Ask which access grants and access policies exist in Vantage, who holds them, and tighten a grant that reaches further than it should.
-
Look for unusual activity in logs
Analyze logs across hosts in Datadog for failed logins, unexpected geographies or spikes from a single source, and search the dashboards your team already built for the same signal.
Things to ask
Ask from ChatGPT, Claude, Cursor or Microsoft Copilot. Each person’s AI can do only what they’re allowed to.
-
Which monitors fired in the last 24 hours?
-
Which action groups have no email receivers in production?
-
Which monitors are alerting right now and on which hosts?
-
Show the Prometheus error rate for checkout over the last hour
-
Which faults in the production project spiked in the last 24 hours?
-
Which spans made checkout slow in the last hour?
-
Find sessions from acme.com this week that hit a checkout error
-
Which alerts fired on the orders table overnight?
For IT and engineering
What an agent can call in Observability
10+ Observability connectors bring 1,439 tools you can call the moment you connect, and 9 of the connectors sign in with nothing to register.
Of those 1,439 tools, 9% delete something. Switch off the delete tools for a role or one person, and their AI can’t use them.
- Read 48%
- 687 tools · list, get, search
- Write 19%
- 279 tools · create, update, send
- Delete 9%
- 136 tools · delete, remove, archive
- Other 23%
- 337 tools · vendor-specific verbs
One endpoint
https://api.elaichi.ai/mcp
Every connector on this page answers here.
Every Observability connector, by depth
Tool counts are what each connector exposes; the split is what those tools do. Hosted means you sign in with nothing to register. Your own app means an admin adds your OAuth app for the connector once, then you connect. API key and App credentials mean an admin pastes them once, with nothing to register.
| Connector | Tools | Read · write · delete | Sign-in |
|---|---|---|---|
| Azure Resource Manager | 521 | Hosted | |
| Datadog | 352 | Hosted | |
| Vantage | 226 | API key | |
| Splunk | 112 | API key | |
| Grafana Cloud | 62 | Hosted | |
| New Relic | 54 | API key | |
| Honeybadger | 39 | Hosted | |
| Honeycomb | 34 | Hosted | |
| Axiom | 31 | Hosted | |
| LogRocket | 8 | Hosted | |
| Monte Carlo | Hosted | ||
| Pydantic Logfire | Hosted |
Questions, answered
How many Observability connectors does Elaichi have?
10+ Observability connectors are in the catalog. Each one arrives as a set of MCP tools an agent can call through Elaichi’s one MCP endpoint.
Can Claude, ChatGPT, Cursor and Microsoft Copilot all use Observability connectors?
Yes. Elaichi exposes one organization-wide endpoint, https://api.elaichi.ai/mcp, and Claude, ChatGPT, Cursor, Microsoft Copilot or any MCP client connects to that same address with OAuth, while the Elaichi Agent reaches the same tools inside the app. Connecting a Observability account once makes it reachable from every one of them.
Do Observability connectors work with Gemini, Codex, Claude Code or other MCP clients?
Yes. Observability connectors are reached over the same MCP endpoint every client uses, so anything that speaks MCP can call them — Gemini, Codex, Claude Code, Windsurf, Cline, Zed and OpenCode among them — alongside Claude, ChatGPT, Cursor, Microsoft Copilot and the Elaichi Agent. There is no per-client setup beyond pointing the client at https://api.elaichi.ai/mcp.
Do Observability connectors need me to bring my own OAuth app?
No. 9 of the Observability connectors use Elaichi's hosted sign-in, with nothing to register. 3 connect with an API key or app credentials that an admin pastes once, with no app to register. Nothing in this category asks you to register an OAuth app before you start.
Can I stop an agent from writing to Observability tools?
Yes. Tool restrictions apply at role and individual level, and a restricted tool can't be used by the people it is restricted for. Read-only access to a Observability connector is a matter of allowing the reads and leaving the writes out.
Put Observability connectors in front of your team
14 days of Black, no credit card. Connect one and pick what each team can call.
When the trial ends, nothing is deleted. You move to the free Personal plan, and connections, roles and the audit log stay where they are.
Add one link to your AI app
The same link for everyone on your team.
https://api.elaichi.ai/mcp
ChatGPT
Claude
Cursor
Microsoft Copilot