Pod

Yes, this is on purpose. Pod is built for agents, so the default page is plain HTML.
Human? View the normal website.
Agent? You probably prefer /mcp/prometheus-mcp-server/issues.md or /mcp/prometheus-mcp-server/issues.json, or Pod over MCP.

Reported issues for Prometheus MCP Server

Pod holds 20 of 29 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 Prometheus MCP Server.

Most discussed

[Feature]: Alert and recording rules tools

Pre-submission Checklist

Feature Type

New MCP Tool (new functionality for AI assistants)

Priority Level

Medium - Would improve workflow significantly

Feature Summary

Tools for managing and interacting with recording and alerting rules.

Read the thread · 2026-05-13 · closed · external user · 3 comments

[Bug]: Helm Chart is not public

Pre-submission Checklist

Priority Level

Medium - Moderate impact on functionality

Bug Description

The helm chart is not available for public.

Steps to Reproduce

  1. Run helm pull oci://ghcr.io/pab1it0/charts/prometheus-mcp-server --version 1.0.0
  2. O

Read the thread · 2026-04-01 · closed · external user · 2 comments

[Bug]: Prometheus mcp server pod not comming 1/1 Running state

Pre-submission Checklist

Priority Level

High - Significant impact, blocks important functionality

Bug Description

Hi @pab1it0

i am trying to install your mcp server with my prometheus url, the configuration looks good !!

<img width="959" height="287" alt

Read the thread · 2026-03-26 · closed · external user · 4 comments

[Question]: How to solve the problem of slow output content

Pre-submission Checklist

Question Type

Configuration Help (setup, environment variables, MCP client config)

Question

When the query time is 30 minutes or longer, the model feedback content will be very slow. Currently, it seems that MCP does not support streaming output. It almost seems like the content is ou

Read the thread · 2025-09-20 · closed · external user · 4 comments

[Bug]: get_metric_metadata

Pre-submission Checklist

Priority Level

Medium - Moderate impact on functionality

Bug Description

Hi,

There is a bug in get_metric_metadata:

https://github.com/pab1it0/prometheus-mcp-server/blob/main/src/prometheus_mcp_server/server.py#L274C18-L274C26

Read the thread · 2025-09-18 · closed · outside contributor · 5 comments

docker run is failed. no response

docker run -it --rm -e PROMETHEUS_URL=http://127.0.0.1:9090/ -e PROMETHEUS_USERNAME= -e PROMETHEUS_PASSWORD= ghcr.io/pab1it0/prometheus-mcp-server:latest [08/19/25 10:29:20] INFO {"source": "environment variables", "note": "No .env file found", "event": "Environment configuration loaded", "level": "info", "timestamp":

Read the thread · 2025-08-19 · closed · external user · 4 comments

Lack of query limit and pagination is problematic in large environments

I'd like to use this more, but our Prometheus is very involved, and several base queries within this tool exceed context limits for Claude Code:

● prometheus-stag101-us-east-1-eks - execute_query (MCP)(query: "kube_pod_info{namespace=\"twodotoh\"}")
  ⎿  Error: MCP tool "execute_query" response (57464 tokens) exceeds maximum allowed tokens (25000). Please use pagination, filtering, or limit parameters to reduce the response size.

● prometheus-stag101-us-east-1-eks - execute_query (

[Read the thread](https://github.com/pab1it0/prometheus-mcp-server/issues/38) · 2025-07-23 · closed · external user · 3 comments

### Docker container doesn't handle `TERM` and `INT` signals and continues to run after mcp server gets stopped in an IDE

## Short description

Each prometheus mcp server restart leaves it's docker container running and orphaned.

## Long description

Use following as an mcp server command configured in the `mcp.json` file:

docker run -i --rm -e PROMETHEUS_URL=http://172.17.0.1:9090 ghcr.io/pab1it0/prometheus-mcp-server:1.0.3


Stop mcp server in the IDE you use and look at the `docker ps`.
The mcp-prometheus docker container continues to run.
Each restart of the mcp server produces one more orphaned runnin

[Read the thread](https://github.com/pab1it0/prometheus-mcp-server/issues/21) · 2025-05-22 · closed · external user · 4 comments

## Most recent

### MCP spec conformance: 9 requirement(s) violated (via @hasmcp/mcp-spec-test)

Running an automated MCP protocol conformance sweep (@hasmcp/mcp-spec-test) against this server found that it does not respond to the protocol handshake (`server/discover` / `initialize`) within the suite's 10-second timeout, even though `PROMETHEUS_URL` was set to a value the server should be able to accept without needing a live, reachable Prometheus instance for the handshake itself. Both spec revisions this server negotiates (2026-07-28 and 2025-11-25) show the same underlying symptom — no r

[Read the thread](https://github.com/pab1it0/prometheus-mcp-server/issues/191) · 2026-08-24 · closed · external user · 1 comment

### [Bug]: get_targets returns all dropped targets unfiltered, exhausting server memory

### Pre-submission Checklist

- [x] I have searched existing issues to ensure this bug hasn't been reported before
- [x] I have checked the documentation and this appears to be a bug, not a configuration issue
- [x] I can reproduce this issue consistently

### Priority Level

High - Significant impact, blocks important functionality

### Bug Description

`get_targets` requests `/api/v1/targets` with no query parameters and returns every dropped target verbatim. Prometheus supports `state` and `s

[Read the thread](https://github.com/pab1it0/prometheus-mcp-server/issues/185) · 2026-08-01 · closed · outside contributor · 1 comment

### [Feature]: add direct Azure Entra ID support to mcp-prometheus as an upstream OIDC provider, without Dex.

### Pre-submission Checklist

- [x] I have searched existing issues and discussions for similar feature requests
- [x] I have checked the documentation to ensure this feature doesn't already exist
- [x] This feature request is related to the Prometheus MCP Server project

### Feature Type

Authentication Enhancement (new auth methods, security)

### Priority Level

High - Important for broader adoption

### Feature Summary

Add direct Azure Entra ID support to mcp-prometheus as an upstream OIDC 

[Read the thread](https://github.com/pab1it0/prometheus-mcp-server/issues/181) · 2026-07-17 · open · external user · 2 comments

### No auth gate on inbound MCP connections (http/sse transport)

Looks like `PROMETHEUS_USERNAME/PASSWORD/TOKEN` authenticate the server's own connection *to* Prometheus, but there's no auth on the inbound side — anyone who can reach the http/sse endpoint can use it.

Flagging in case it's useful: [getmcpauth](https://getmcpauth.dev) does OAuth 2.1 + DCR specifically for MCP servers, could close that gap without a big refactor. Not trying to sell hard, just noticed it while looking at MCP servers with exposed HTTP transports.

[Read the thread](https://github.com/pab1it0/prometheus-mcp-server/issues/177) · 2026-07-01 · closed · external user · 1 comment

### [Bug]: helm chart mentioned in the docs seems to be private

### Pre-submission Checklist

- [x] I have searched existing issues to ensure this bug hasn't been reported before
- [x] I have checked the documentation and this appears to be a bug, not a configuration issue
- [x] I can reproduce this issue consistently

### Priority Level

Low - Minor issue, workaround available

### Bug Description

helm chart mentioned in the docs seems to be set to private

### Steps to Reproduce

helm pull oci://ghcr.io/pab1it0/charts/prometheus-mcp-server Error: G

Read the thread · 2026-03-12 · closed · external user · 2 comments

[Bug]: 严重的性能问题，只要打开IDE，MCP就会一直占用CPU！！！

Pre-submission Checklist

Priority Level

Medium - Moderate impact on functionality

Bug Description

代码里其实写了一套缓存逻辑（get_cached_metrics），但开发者在最重要的 list_metrics 工具里完全忘了去调用它。导致每一次 IDE 检查MCP插件时，它都要从 Prometheus 重新拉取并解析上万条指标数据。会持续占用CPU。

Steps to Reproduce

  1. 只要

Read the thread · 2026-02-06 · closed · external user · 2 comments

[Feature]: Support for Multiple MCP Servers in Cursor

Pre-submission Checklist

Feature Type

New MCP Tool (new functionality for AI assistants)

Priority Level

High - Important for broader adoption

Feature Summary

When attempting to run multiple instances of the MCP server in Cursor (to sup

Read the thread · 2026-01-09 · closed · outside contributor · 1 comment

[Question]: Deployment is complete, but no ports are exposed. How can I verify this?

Pre-submission Checklist

Question Type

Configuration Help (setup, environment variables, MCP client config)

Question

Image

Context and Background

服务部署完成，但没有监听的端口，如何验证和使用呢

Read the thread · 2025-12-09 · closed · external user · 1 comment

[Bug]: Failed to validate tool mcp_prometheusmcp_delete_series:

Pre-submission Checklist

Priority Level

High - Significant impact, blocks important functionality

Bug Description

docker run --rm -p 8080:8080 ghcr.io/tjhop/prometheus-mcp-server:latest --prometheus.url "http://localhost:9090" --mcp.transport "http" --web.l

Read the thread · 2025-10-29 · closed · external user · 2 comments

[Bug]: get_metric_metadata needs to use a different endpoint

Pre-submission Checklist

Priority Level

Low - Minor issue, workaround available

Bug Description

Currently, get_metric_metadata is implemented using 'metadata' endpoint and params like {"metric": } https://github.com/pab1it0/prometheus-mcp-server/blo

Read the thread · 2025-10-27 · closed · outside contributor · 1 comment

[Feature]: Adding Option to Disable SSL verification of the Prometheus URL for Test Environments

Pre-submission Checklist

Feature Type

Configuration Option (new settings, customization)

Priority Level

Medium - Would improve workflow significantly

Feature Summary

PR: https://github.com/pab1it0/prometheus-mcp-server/pull/84

This fe

Read the thread · 2025-10-25 · closed · outside contributor · 1 comment

[Feature]: Would it be useful to add a generic discovery endpoint?

Pre-submission Checklist

Feature Type

New MCP Tool (new functionality for AI assistants)

Priority Level

Medium - Would improve workflow significantly

Feature Summary

If we have a discovery endpoint, the LLMs can get an understanding of

Read the thread · 2025-10-21 · closed · external user · 1 comment

The remaining reports are on the project's issue tracker.