Add Temporal Proxy dashboard

[gauravthadani]

Summary

• Adds a new Grafana dashboard at proxy/proxy.json covering the metrics emitted by temporal-proxy (server / router / encryption subsystems, per the observability docs).
• Adds proxy/ to the top-level README's Directory structure and links the proxy observability page under Available metrics.

Layout

• Overview: success rate, total RPS, error rate (excludes Canceled and NotFound — client-driven codes that aren't proxy failures).
• Traffic: RPS by method (short-legend via label_replace to strip the gRPC service prefix), RPS by gRPC code (stacked, OK colored green).
• Latency: p50 / p95 / p99 request duration by method.
• Router: routing decisions by upstream/outcome; forwarding errors by upstream/reason.
• Encryption: vault ops, KEK ops (by provider), DEK ops + rotations, DEK cache hit ratio.

Two template variables: datasource (Prometheus, matches the convention used in cloud/temporal_cloud.json) and upstream.

Test plan

• Import proxy/proxy.json into Grafana, pick the Prometheus datasource for the temporal-proxy metrics, and confirm every panel renders.
• Confirm RPS by method legends show short RPC names (e.g. StartWorkflowExecution) rather than the full /temporal.api.workflowservice.v1.WorkflowService/StartWorkflowExecution.
• Confirm the Error rate stat stays low when traffic is dominated by long-poll Canceled / lookup NotFound results.

🤖 Generated with Claude Code