Skip to content

Hermes Gateway Monitoring — Prometheus Metrics & Cache Reports

Hermes Gateway provides two complementary observability interfaces: a Prometheus metrics endpoint for real-time monitoring, and an in-proxy SQL console for cache analysis and query diagnostics.

For most operational use cases, Prometheus metrics are the right tool. The SQL console is better suited for ad-hoc query analysis and debugging.

This repo includes a Docker Compose setup that brings up Grafana and Prometheus together with a pre-built sample dashboard. See the Installation Guide for setup details.

Prometheus Metrics

Hermes Gateway exposes the following Prometheus metrics:

Cache Metrics - cacheQueries — total queries handled by the cache layer - cacheHits / cacheMisses / cacheExpires / cacheLocks — per-table cache performance (labelled by table) - totalCacheHits / totalCacheMisses / totalCacheLocks — aggregate cache totals

Query Metrics - queryDuration — per-table query execution time - queryRows — per-table row counts - totalQueryDuration / totalQueryRows — aggregate query totals since last restart - totalQueryErrors — cumulative query error count

Connection Metrics - clientConnectionsTotal / clientConnectionsCurrent — connections from applications to Hermes Gateway - serverConnectionsTotal / serverConnectionsErrors / serverConnectionsCurrent — connections from Hermes Gateway to the upstream database

Usage Metrics - tableUpdates — upstream table update events detected - sentTableUpdates — update events forwarded to downstream consumers

Cache Reports

Hermes Gateway's cache reporting surfaces actionable insights about query performance and cache efficiency — helping teams identify optimisation opportunities without instrumenting application code.

Currently available:

  • Identify the slowest queries and the tables they involve
  • Identify queries against tables without indexes
  • Identify queries with the highest cache hit rates, showing where the proxy is delivering the most performance uplift

Grafana Dashboard

A sample Grafana dashboard is available in the monitoring/grafana folder. Import it into your Grafana instance after configuring Prometheus to scrape the Hermes Gateway metrics endpoint.

Logging

Hermes Gateway logs startup, shutdown, and error events to stdout, available via standard Docker log tooling. Log verbosity and output destination are configured via the log section — see the Configuration Guide for details.