[FEATURE] Add configurable prometheus.url in values.yaml with kubectl auto-discovery fallback

[AbdulazizTamkeen]

📋 Prerequisites

• I have searched the existing issues to avoid creating a duplicate
• By submitting this issue, you agree to follow our Code of Conduct

📝 Feature Summary

Add prometheus.url field to observability-agent's values.yaml, templated into the system prompt, with kubectl-based auto-discovery as a fallback

❓ Problem Statement / Motivation

• What is the current limitation or pain point?
  The only way to give observability-agent the Prometheus URL is by manually editing the system prompt at runtime. Every time the agent restarts (e.g. after a Helm upgrade or pod restart), it reverts to the default system prompt from the chart, losing the custom URL.

• Who is affected?
  Anyone running observability-agent alongside an in-cluster Prometheus instance.

• Why is this needed?
  Without a persistent way to configure the Prometheus URL, users must re-edit the system prompt after every restart, which is error-prone and not GitOps-friendly.

💡 Proposed Solution

• What should be added/changed?

  1. Add a prometheus.url field to observability-agent's values.yaml
  2. Template it into the agent's system message, similar to how grafana-mcp handles grafana.url

• How should it work?
  If prometheus.url is set in values.yaml, it is injected into the system prompt at deploy time and persists across restarts.

• Any specific requirements or constraints?
  N/A

🔄 Alternatives Considered

• Manually editing the system prompt at runtime (current workaround), not persistent across restarts

🎯 Affected Service(s)

App Service

📚 Additional Context

N/A

🙋 Are you willing to contribute?

• I am willing to submit a PR for this feature