40901d8ad3
All acceptance criteria already met: - Sizing table reproduced from plan §14.7 - Redis memory accounting paragraph included - Worked example for ≤200 GB tier - Links from README.md and production.md The sizing guide is THE artifact operators need on day one. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
1.4 KiB
1.4 KiB
P6.9: Deployment Sizing Matrix (bf-7r59)
Status: Already Complete
All acceptance criteria verified:
- Sizing table reproduced -
docs/horizontal-scaling/sizing.mdcontains the 5-row matrix from plan §14.7 - Redis memory accounting - Paragraph documenting ~20 MB per 10k active IPs for search UI rate-limit buckets
- Worked example - ≤ 200 GB / ≤ 5 k QPS tier with memory budget (§14.2) and QPS (§14.3) validation
- README.md link - Line 89 links to sizing guide in "Production deployment" subsection
- production.md link - Lines 7, 48, and 186 link to sizing guide
Content Summary
The sizing matrix provides operators with:
| Corpus | Peak QPS | Orchestrator pods | Task store |
|---|---|---|---|
| ≤ 10 GB | ≤ 500 | 2 (HA) | Redis (or SQLite if replicas=1) |
| ≤ 50 GB | ≤ 2 k | 2–4 (HPA) | Redis |
| ≤ 200 GB | ≤ 5 k | 4–8 (HPA) | Redis |
| ≤ 1 TB | ≤ 20 k | 8–12 (HPA) | Redis |
| ≤ 5 TB | ≤ 100 k | 12–24 (HPA) | Redis (clustered or Sentinel) |
Key insight: Orchestrator count scales with query throughput; Meilisearch node count scales with corpus size and replication factor. These are orthogonal axes.
Redis backs shared state for idempotency, session pinning, alias cache, job queue, leader lease, CDC overflow, and search UI rate-limit buckets (~20 MB per 10k active IPs).