Health

The Health resource provides service health monitoring and system metrics for your Glow instance. Use it to track uptime, latency, CPU, memory, and error rates over time. Health is an API-only resource with no CLI commands.

What is Health?

Health aggregates hourly service check results and system metrics into time-series views. It answers questions like: Is the instance healthy? What was the uptime last week? Are latency or error rates trending up? Administrators use health data to monitor their university’s Glow deployment and diagnose performance issues before they affect users.

The health response includes two materialized views:

service_hourly — per-hour uptime, check counts, latency, and error status for each monitored service.
metrics_hourly — per-hour CPU, memory, request counts, and error totals.

Quick Start

API

Fetch health data for the last 7 days:

Calls below use $GLOW_INSTANCE_URL + $GLOW_TOKEN — see Authentication to export them once.


curl -X POST $GLOW_INSTANCE_URL/system/health \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "date_from": "2026-03-21",
    "date_to": "2026-03-28",
    "page_limit": 100,
    "page_offset": 0
  }'

Filter by service:


curl -X POST $GLOW_INSTANCE_URL/system/health \
  -H "Authorization: Bearer $GLOW_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "service": "glow-api",
    "date_from": "2026-03-21",
    "date_to": "2026-03-28"
  }'

Understanding the Response

Service Hourly View

Each service_hourly entry contains:

Field	Description
`date_hour`	Hour bucket (e.g., `"2026-03-28T14:00:00Z"`)
`service`	Service name (e.g., `"glow-api"`)
`check_count`	Total health checks in this hour
`ok_count`	Successful checks
`fail_count`	Failed checks
`uptime_percent`	Uptime as a percentage (0—100)
`avg_latency_ms`	Average response latency in milliseconds
`min_latency_ms`	Minimum latency observed
`max_latency_ms`	Maximum latency observed
`latest_ok`	Whether the most recent check passed
`latest_error`	Error message from the most recent failure (empty if healthy)

Metrics Hourly View

Each metrics_hourly entry contains:

Field	Description
`date_hour`	Hour bucket
`sample_count`	Number of metric samples collected
`avg_cpu_percent` / `min` / `max`	CPU utilization
`avg_latency_ms` / `min` / `max`	Request latency
`avg_memory_bytes` / `min` / `max`	Memory usage
`max_requests_total`	Peak total request count
`max_errors_total`	Peak total error count

Refreshing and Exporting

Refresh materialized views to ensure health data is current:


curl -X POST $GLOW_INSTANCE_URL/system/refresh \
  -H "Authorization: Bearer $GLOW_TOKEN"

Returns success, refreshed_views, and invalidated_tags.

Export health data as a denormalized ZIP for offline analysis:


curl -X POST $GLOW_INSTANCE_URL/system/export \
  -H "Authorization: Bearer $GLOW_TOKEN"

Returns content (base64-encoded), file_name, mime_type, and row_count.

Common Operations

Operation	Method	Endpoint
Get health data	`POST`	`POST /system/health`
Refresh views	`POST`	`POST /system/refresh`
Export data (ZIP)	`POST`	`POST /system/export`
Get documentation	`POST`	`/health/docs`

Health API
Info Guide — quick server status check