Hello, I’m Nora, your AI guide from Technology Last Stop. This guide covers Run observe and recover storage collection. Use owner-only measurements and explicit scheduler evidence for capacity decisions. Current availability: Implemented; verification required.

Before you start. An authorized administrator in the correct company or the TLS owner console, with mandatory verification completed. Use an approved synthetic account for acceptance. Keep credentials, personal data and recovery codes out of recordings.

Step 1. TLS owners open Storage & capacity. Current records live in the environment’s shared application database, separated by company ID. File objects use protected object storage. A customer does not currently receive a physical database just by subscribing.

Step 2. Record an actual provider allocation with its reference and evidence. A pool without verified allocation or a complete recent reading remains unknown. Company file metadata does not include all database pages, backups, untracked objects or provider charges.

Step 3. Run a manual collection to confirm access. Database and object observations are separate. Large object inventories resume from a durable cursor; an unfinished pass does not claim a complete total.

Step 4. Configure the independent runner using the owner-only endpoint and a protected 90-day runner token. The downloadable Node 22 runner includes the contract. Store its token in the scheduler’s secret manager; never in a URL or source repository.

Step 5. Accept a real scheduled invocation and inspect the recorded caller, last started/completed time, lease and error. In-app traffic recovery is not proof of cron operation. The backup capacity watch only reads observations and cannot start collectors.

Step 6. Use default 75% warning and 90% critical thresholds or enter an approved pool-specific pair. Investigate newly stale/error readings as well as capacity thresholds. Acknowledgment records review; it does not clear the measured condition.

Step 7. For expansion, review costs, margin, backup/restore and a provider capacity-change proposal. Approve the provider purchase separately, verify the resulting allocation and fresh measurement, then record completion. Changing a number in this console does not purchase storage.

Step 8. Pause, repair and rotate the runner when needed. Interrupted work can resume after the bounded lease expires. Restore recycled component records within 60 days; deleting their inventory record does not delete physical infrastructure.

Now verify the result. A stale or unverified pool cannot be called healthy. Collectors and alerts are visible only to TLS owners. The same running collection cannot be claimed by overlapping invocations.

If you get stuck. Provider quotas, cloud allocation credentials, independent scheduling and email receipt still require live acceptance. No automatic purchase or overage charge is triggered by capacity thresholds.

Use the current written guide for exact URLs, permission names, and provider documentation. Pause before entering credentials, and keep them in protected configuration. I’m Nora, your TLS guide.