# Reported issues for Dynatrace-mcp

Pod holds 23 of 41 problems reported by people outside the maintainer team. Issues filed by the project's own owners, members and collaborators are excluded entirely — a maintainer's release checklist is not a warning to a prospective user.

Back to [Dynatrace-mcp](/mcp/dynatrace-mcp).

## Most discussed

### Unable to connect to Dynatrace when deploying mcp server on EKS which connects through AWS private link

I am trying to deploy Dynatrace MCP server on AWS EKS which has private link connectivity to multiple Dynatrace tenants.

I am using platform token with apps.dynatarce.com endpoints as mentioned on github repo. 

Would this MCP server runs in this type of scenario?
I'll be building custom MCP client frontend and backend which connects to Dynatrace MCP server and uses LLM for below uses cases  - 

Real-time observability - Fetch production-level data for early detection and proactive monitoring

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/199) · 2025-10-13 · closed · external user · 9 comments

### Response in HTTP format

**Describe the bug**
When using the MCP server in HTTP mode, it responds back in SSE format instead of HTTP/JSON

**To Reproduce**
Steps to reproduce the behavior:
Connect to the MCP via a client: MCP Inspector, Cline, simple curl, etc
List tools
Get problems

**Expected behavior**
Get an HTTP/JSON format response. SSE is returned instead

**Screenshots**

**Desktop (please complete the following information):**

- OS: [e.g. iOS]
- Browser [e.g. chrome, safari]
- Version [e.g. 22]

**Smartphone

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/144) · 2025-09-10 · closed · external user · 5 comments

### Unable to connect to the MCP server using Langchain MultiServerMCPClient

**Describe the bug**
When running the Dynatrace MCP server in http mode, the MultiServerMCPClient (python) fails to connect to the mcp server and retrieve list of tools. My MCP server is running on IP address 52.186.168.XXX on port 3000. I have obfuscated the last part of the IP address for obvious reasons.

**To Reproduce**
Steps to reproduce the behavior:
This is the snippet of code from my Agentic AI app
```
                if not self.config_manager.validate_config(mcp_config):
             

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/133) · 2025-09-03 · closed · external user · 5 comments

### Unable to access

I am trying to use your code like creating .env file that include client id , client secret and environment but i am getting error like trying to run my file but instead of json response i am getting html page code and getting error like 

Trying to authenticate API Calis to https://rhw1968.live.dynatrace.com/ via OAuthClientId dt0s02. 2WSDFNOZ
Using SSO auth URL: https://rhw1968.live.dynatrace.com/sso/oauth2/token
Fatal error in main(): SyntaxError: Unexpected token °<°,

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/31) · 2025-06-16 · closed · external user · 6 comments

### Expired Dynatrace MCP token error message

**Describe the bug**
One day after I installed the Dynatrace MCP server on my machine, I started getting errors like this:

```bash
Your Dynatrace MCP authentication token has expired, so I cannot retrieve the list of services at this time. Please refresh your credentials in the s.env.av file and restart the MCP server. Once re-authenticated, I can fetch and display the list of services for you.
Let me know when you've updated your credentials!
```

My OAuth configurations were correct and worke

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/27) · 2025-06-11 · closed · external user · 5 comments

### VSCode never uses the MCP Servers

**Describe the bug**
MCP Server starts, but copilot never uses the MCP server. Any prompts about the exposed tools bring nothing from the MCP serverr.

This is the output:
2025-05-20 16:29:32.233 [info] Connection state: Starting
2025-05-20 16:29:32.236 [info] Starting server from LocalProcess extension host
2025-05-20 16:29:32.237 [info] Connection state: Starting
2025-05-20 16:29:32.238 [info] Connection state: Running
2025-05-20 16:29:37.239 [info] Waiting for server to respond to `initialize

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/21) · 2025-05-20 · closed · external user · 10 comments

### Difference in attributes for filtering

I have noticed difference in the `dql` query used to filter logs between the one mentioned in mcp server (`fetch logs | filter dt.source_entity == "${entityId}"`) and one used in my org (`fetch logs | filter matchesValue(service,"${entityId}")`). I have checked attributes returned from the `fetch logs` query in our instance, no attribute named `dt.source_entity` is being returned due to which tool `get-logs-for-entity` is returning an empty array. Is the difference in the query expected? if yes 

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/13) · 2025-05-11 · closed · external user · 5 comments

### is DT Managed supported?

Do you plan to support DT managed?

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/10) · 2025-05-07 · closed · external user · 10 comments

## Most recent

### Want to add your plugin to awesome-ai-plugins?

Hey, we maintain [awesome-ai-plugins](https://github.com/hashgraph-online/awesome-ai-plugins) and came across `dynatrace-mcp`. It looks like a great fit alongside the other gemini cli extensions, and we'd love to include it.

Would you be up for submitting it? The whole submission is a small PR adding one entry to `README.md`; [`CONTRIBUTING.md`](https://github.com/hashgraph-online/awesome-ai-plugins/blob/main/CONTRIBUTING.md) shows the format.

Projects listed there can also receive a public pr

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/572) · 2026-08-26 · closed · external user · 1 comment

### OAuth token not persisted across server restarts — breaks MCP clients that restart the process

## Description
  
  MCP clients like Kiro CLI spawn a new `mcp-server-dynatrace` process per session. Since the OAuth token cache is in-memory only (`InMemoryTokenCache`), every new
  session triggers a browser authorization flow. This makes the server unusable for any MCP client that doesn't keep the process alive indefinitely.
  
  ## Steps to Reproduce
  
  1. Configure `mcp-server-dynatrace` with `DT_ENVIRONMENT` only (OAuth auth code flow)
  2. Start the server — browser opens, authorize su

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/542) · 2026-06-14 · closed · external user · 1 comment

### CLI --help and --version require DT_ENVIRONMENT before printing metadata

## Summary

The published `@dynatrace-oss/dynatrace-mcp-server` CLI starts environment validation before handling top-level metadata flags.

Without `DT_ENVIRONMENT`, `mcp-server-dynatrace --help` and `mcp-server-dynatrace --version` both fail with the same server startup error as a normal launch. Users cannot inspect usage or the installed version unless they already have Dynatrace environment variables configured.

## Reproduction

Tested on Windows with Node.js 24.16.0 and `@dynatrace-oss/dyn

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/523) · 2026-06-07 · closed · outside contributor · 0 comments

### Extensions support

**Is your feature request related to a problem? Please describe.**
Currently I'm unable to manage extenions using mcp serverand I have to use dtctl

**Describe the solution you'd like**
I want to be able to manage extensions using mcp server

**Describe alternatives you've considered**
dtctl

**Additional context**

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/495) · 2026-04-28 · closed · external user · 1 comment

### Prevent reauth on every session

**Is your feature request related to a problem? Please describe.**
Auth window opens Everytime I open a new instance of the coding agent .

I also have 5 different instances of the mcp server configured (each for a different dt instance) 

**Describe the solution you'd like**
Save authentication token in a file inside the home directory (or wherever you feel appropriate) and only open the browser auth window once a day (or whenever that auth expired)

Thanks

**Describe alternatives you've consi

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/485) · 2026-04-21 · closed · external user · 1 comment

### Claude Code MCP Connector is not connecting

**The bug**
I tried to add the DT MCP to Claude code desktop app. The options listed for connecting the MCP included a required field for the tenant — the rest was optional. 

**To Reproduce**
Steps to reproduce the behavior:

1. Go to Claude Desktop
2. Click on Connectors / Browse connectors 
3. Install DT MCP 
4. Fill the form but leave the optional fields empty 
5. enable it 
6.. See the error

<img width="560" height="229" alt="Image" src="https://github.com/user-attachments/assets/d7ce1a6d-

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/479) · 2026-04-19 · closed · external user · 3 comments

### Server crashes immediately after initialize handshake following Claude desktop update (April 2026)

```markdown
## Summary

Following a Claude desktop app update on April 17, 2026, the Dynatrace MCP Server (v1.7.3) consistently fails to connect. The server process crashes within ~70ms of receiving the `initialize` message, before it can send any response. The server was working correctly earlier the same day.

## Environment

- Dynatrace MCP Server version: 1.7.3
- Client: Claude desktop (claude-ai v0.1.0)
- MCP Protocol version: 2025-11-25
- Dynatrace environment: `https://<your-tenant>.apps.

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/478) · 2026-04-17 · closed · external user · 3 comments

### Issue with dt.system_events query

Hi team,

I posted this internally on Slack already but the DQL query that we provide for users to track the queries made by the MCP server is incorrect after a change made to the semantic dictionary. The client context field has changed, and the value inside of that field slightly changed. Please see the original and fixed versions below:

The original query:

`fetch dt.system.events
| filter event.kind == "QUERY_EXECUTION_EVENT" and contains(client.client_context, "dynatrace-mcp")
| sort times

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/420) · 2026-03-24 · closed · external user · 1 comment

### chat_with_davis_copilot: Remove repeating LLM question

**Is your feature request related to a problem? Please describe.**
Refer to: https://github.com/dynatrace-oss/dynatrace-mcp/blob/main/src/index.ts#L1001

Waste of context space? (feedback from customers testing MCP)

**Describe the solution you'd like**

Remove redundant question text.

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/340) · 2026-02-23 · closed · external user · 0 comments

### Feature Request: Multi-User Authentication Support for Centralized (OpenShift/Kubernetes) Deployments

**Is your feature request related to a problem? Please describe.**
We are deploying the Dynatrace MCP Server on **OpenShift/Kubernetes** (`--http` mode) for **multiple users**. Currently all users share the same credentials — so there is no per-user identity, no per-user RBAC, and no per-user audit trail.

**Describe the solution you'd like**
**What we need:** Per-user authentication support so each user connects with their own Dynatrace identity/token.

**Questions:**
1. Is multi-user support p

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/328) · 2026-02-16 · closed · external user · 2 comments

### Multi‑Tenant Support for Dynatrace MCP Server

**Is your feature request related to a problem? Please describe. **
The Dynatrace MCP Server currently supports only one Dynatrace tenant per instance.

We are a large financial enterprise using **15+ Dynatrace SaaS tenants**, and maintaining separate MCP instances for each tenant creates unnecessary operational overhead, complexity, and maintenance challenges.

**Describe the solution you'd like**

1. Native multi‑tenant support within a single MCP Server instance, including:
2. Ability to conf

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/303) · 2026-01-22 · closed · external user · 3 comments

### Unable to start the Dynatrace CMP HTTP server in Docker

**Describe the bug**
A clear and concise description of what the bug is.

**To Reproduce**
Steps to reproduce the behavior:

1.  Pull the Dynatrace MCP container from https://hub.docker.com/r/mcp/dynatrace-mcp-server
2.  Start it
3.  The Dynatrace MCP container would only start as stdio mode.  How to pass --http parameter to the container?

**Expected behavior**
Dynatrace MCP container can support both stdio and http

**Screenshots**
If applicable, add screenshots to help explain your problem.

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/288) · 2025-12-24 · closed · external user · 5 comments

### DQL generated by nl2dql is incorrect

**Describe the bug**
When using the natural language to get more details about a specific problem, the generated DQL is incorrect. In the DQL, it is filtering the query on field event.problem_id. However, it should be filtering on display_id. When I run the DQL generated by nl2dql in the notebook, I don't get any data in my tenant. Upon further investigation, it appears that event.problem_id is not a field anymore.

**To Reproduce**
Steps to reproduce the behavior:
1. Use this nl to query all pr

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/284) · 2025-12-17 · closed · external user · 3 comments

### Does it work in DynTrace managed mode?

Hey team, 

As the title suggests, I want to know if this works in both Dynatrace SaaS and Dynatrace Managed modes.

Thanks

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/280) · 2025-12-15 · closed · external user · 1 comment

### Failing integration tests

When I checkout the repo and run `npm test`, some of the integration tests fail.

See PR https://github.com/dynatrace-oss/dynatrace-mcp/pull/220 for a fix.

- find-monitored-entity:
    - Uses findMonitoredEntityByName, which no longer exists
    - Assertions assume result is a string, but it is a DqlExecutionResult
    - Test for blank entity "" fails.
- send-emails:
    - auth integration test fails, because exception is now thrown by createDtHttpClient rather than sendEmail.
- davis-copilot-e

[Read the thread](https://github.com/dynatrace-oss/dynatrace-mcp/issues/221) · 2025-10-22 · closed · outside contributor · 1 comment

The remaining reports are on [the project's issue tracker](https://github.com/dynatrace-oss/Dynatrace-mcp/issues).
