ops: Disk tier occupancy on /metrics
CI / vulncheck (pull_request) Successful in 19s
CI / check-and-test (pull_request) Successful in 47s
CI / vulncheck (push) Successful in 16s
CI / check-and-test (push) Successful in 46s
Release Tag / release (push) Successful in 14s

Operators could see attach-ready and capacity-pressure events but not how
full the configured disk (or memory) tier was without reading filesystems.
Expose size/capacity gauges and disk_cache_full_ratio next to disk_tier_ready.
Capacity is a config read, so it stays available while attach is pending.
This commit was merged in pull request #53.
This commit is contained in:
2026-09-09 16:06:12 +00:00
parent a3ea4806a7
commit 50ca0a071c
5 changed files with 240 additions and 2 deletions
+4
View File
@@ -88,6 +88,9 @@ curl -s -i http://localhost/lancache-heartbeat
| `total_requests` / `errors` | Volume and failures |
| `upstream_errors` / `cache_write_failures` / `rate_limited` | Upstream pipe, cache write, and rate-limit pressure (Quick check highlights these next to hit/miss) |
| `disk_tier_ready` | `0` while disk slow-tier attach pending; `1` when attached, or when no disk configured (N/A — not waiting) |
| `memory_cache_size` / `disk_cache_size` | Current cache occupancy per tier (bytes) |
| `memory_cache_capacity` / `disk_cache_capacity` | Configured capacity per tier (bytes); `disk_cache_capacity` is `0` when no disk is configured |
| `disk_cache_full_ratio` | `disk_cache_size / disk_cache_capacity` in [0,1]; 0 when no disk is configured or capacity is 0. Tells you "95% full" vs "barely filled" without reading the filesystem |
| `capacity_pressure_events` | Soft eviction under the memory or disk cap, and/or disk Create/Write/Mkdir hitting ENOSPC (volume full). Distinct from cold-cache misses and from the existing `evictions` counter. Logs `tier` (memory or disk) and `reason` (eviction or enospc). |
A first pass through new content is mostly misses (`hit_rate` near 0). Repeat the same content and `cache_hits` / `hit_rate` should rise.
@@ -311,6 +314,7 @@ See `config.Validate()` and `steamcache.New` error paths. This ensures the LAN a
- Godoc on `disk.New` and `DiskFS.Size` expanded with the barrier/attach behavior.
- Startup logs: Info "Disk slow tier attach pending..." then later "Disk slow tier attached (...)" for disk-only and mixed modes.
- `/metrics` exposes `disk_tier_ready` 0/1 and stays responsive during attach (GetMetrics does not block on Size while pending).
- `/metrics` tier occupancy: `memory_cache_size` / `disk_cache_size` (bytes in use) next to `memory_cache_capacity` / `disk_cache_capacity` (configured capacity; `disk_cache_capacity` is 0 when no disk is configured), plus `disk_cache_full_ratio` (size/capacity in [0,1]). Capacity is a config read, so it is reported even while the disk attach is pending (size stays 0 until attach).
- `/lancache-heartbeat` header `X-SteamCache-Disk-Tier` mirrors that state.
- `/metrics` `capacity_pressure_events` counts times the cache dropped data under capacity pressure (soft eviction at the memory or disk cap, or disk Create/Write/Mkdir returning ENOSPC). Logs include `tier=memory|disk` and `reason=eviction|enospc` so operators can grep and tell this apart from a cold cache. The existing `evictions` counter is unchanged.