400 Error monitoring tab some services

[StefanVDWeide]

To Reproduce

1. Create Nuxt 4 service
2. Go to monitoring tab

Current vs. Expected behavior

I run 6 different projects on my self-hosted (hetzner) dokploy instance and for 3 of them I have never been able to see the monitoring tab for the services inside of them. For some reason I always get a 400 error with the following console output:

framework-5115b136cd71a45c.js:1 TypeError: c.cpu.value.replace is not a function
    at A (6798-2a5d0bc8aab0eb2b.js:5:12660)
    at ak (framework-5115b136cd71a45c.js:1:68190)
    at up (framework-5115b136cd71a45c.js:1:80041)
    at i (framework-5115b136cd71a45c.js:1:127501)
    at oD (framework-5115b136cd71a45c.js:1:106283)
    at oO (framework-5115b136cd71a45c.js:1:106153)
    at oE (framework-5115b136cd71a45c.js:1:102924)
    at ox (framework-5115b136cd71a45c.js:1:101479)
    at x (framework-5115b136cd71a45c.js:1:137396)
    at T (framework-5115b136cd71a45c.js:1:137917)

I've tried redeploying the services but that unfortunately didn't solve the issue.

The services that keep crashing are simple Nuxt websites and one Python Flask API.

Is there something that can be changed and/or updated to resolve this? It would be great to have monitoring for all projects.

Provide environment information

I am running:

Provider: `Hetzner`
OS: `Ubuntu 24.04 LTS`
Version: `v0.28.8`
Engine Version: `27.1.2`

Which area(s) are affected? (Select all that apply)

Application

Are you deploying the applications where Dokploy is installed or on a remote server?

Same server where Dokploy is installed

Additional context

No response

Will you send a PR to fix it?

No

[dosubot]

Hi @StefanVDWeide! I'm Dosu and I’m helping the dokploy team.

This is a known issue caused by a type mismatch in the monitoring data pipeline. The frontend expects cpu.value to be a string and calls .replace("%", "") on it, but the backend storage function accepts number | string | unknown, allowing numbers to slip through during JSON serialization.

This affects some services but not others due to how Docker stats get parsed and stored per container.

Similar reports: Issue #1235 ("m.cpu.value is null - happens sporadically with some applications") and Issue #464 (400 errors on monitoring tab).

Solution: PR #3078 was merged in December 2025 to fix CPU value parsing. Since you're on v0.28.8 (likely before this fix), upgrading to the latest Dokploy version should resolve the issue.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[StefanVDWeide]

@dosu version 0.28.8 was the latest version where this was supposed to be fixed already. Hence me opening this issue.

[dosubot]

You're right that PR #3078 is included in v0.28.8. However, looking at the code, that fix was incomplete — it updated the TypeScript type definition to expect cpu.value as a string, but it didn't add runtime validation before calling .replace() [1].

Interestingly, memory values have defensive type checking (if (!memoryString || typeof memoryString !== \"string\")), but CPU values don't have the same protection. So when the WebSocket sends a non-string CPU value (number, null, etc.) during certain container states, the .replace() call fails.

This looks like a bug that needs a proper runtime type guard similar to how memory is handled. Thank you for reporting it!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}