Documentation ›Operations
OPERATIONS / SUMMARY

Objectives, capacity & recovery

What historical evidence establishes and which readiness gaps remain.

Goals and trends

Applications use a provisional 99.9% availability goal over 30 days unless an existing objective owns the signal. The page shows observed coverage and error-budget use. A short history is clearly provisional. Missing latency or freshness measurements remain visible gaps.

Plan capacity

Daily capacity estimates use up to 14 days of history. At least seven days and 80% coverage are required. Disk exhaustion dates appear only for a credible growing trend. CPU, memory, and queue observations appear separately.

Know what can recover

Recovery readiness lists backup scope, result, age, and verified tests where available. A successful integrity check is not the same as a complete application recovery test. Missing backup jobs stay highlighted; this release does not restore production automatically.

Fantasy inference performance

Fantasy Qwen must complete real chat requests successfully before the application’s existing 220-second deadline. At least five completions in the last hour and one in the last fifteen minutes are required. Health probes and isolated recovery tests do not count. Missing activity stays unverified.

ON THIS PAGE

THREE WAYS TO UNDERSTAND

Same system.
Your level of detail.

Switch reading modes above to find the explanation that fits.

Explore Expert →

Keep exploring

Public observability workspace

Public native dashboards, server navigation and owner changes.

↗
Backups & recovery

Protect existing data and know which recovery path to use.

↗