Operations and recovery

Operations helps administrators review safe system health signals, check maintenance tools, and retry supported failed work without exposing private content, provider secrets, or raw storage details.

Open Operations

1. Sign in with an administrator account.
2. Open Admin > Dashboard.
3. Use Overview, Readiness, Maintenance, and Jobs to move between the available operation groups.
4. Select Refresh to reload the latest operation summaries before taking action.

If you are not an administrator, CareMap shows Admin access is required and does not load the operations panels. In development builds, Admin > Development also shows the redacted Release Gate diagnostics used before a release.

Read the status first

The Operations overview summarizes failed, retryable, running, pending, and completed work. Start there before retrying anything so you can separate active jobs from work that needs attention.

The Readiness group includes:

• Backup Readiness for redacted data and storage manifest checks.
• Readiness for provider, route, and worker signals.
• CareMap RAG for queue, worker, artifact, rebuild, and accounting health.
• Knowledge Base for release-local Help content, active search corpus, and documentation ingestion state.

Readiness labels can include Ready, Missing Config, Degraded, Unknown, or Loading. Open a readiness chip to review check details, missing configuration, and recent operation signals. RAG and Knowledge Base summaries omit content and sensitive identifiers.

Use maintenance tools carefully

Maintenance actions are designed to show safe summaries before or after they run:

• Backup Readiness can refresh the manifest and run a restore preflight. The preflight reports whether restore checks are ready, warning, blocked, not applicable, or not checked; it does not restore data.
• Asset GC can scan for unreferenced storage objects and stale authoring images before deletion. Results use redacted paths and summarize garbage counts, stale authoring images, deleted items, and reclaimed size.
• Notifications can test delivery readiness, dry-run stale subscription cleanup, revoke stale subscriptions, or delete already revoked subscriptions. Results use safe provider, channel, destination, and hashed endpoint labels; they do not expose notification tokens, endpoints, or provider payloads.
• Knowledge Base can request synchronization again when the current state is eligible for retry.

Use the visible result and reason text to decide whether a follow-up action is safe. If a control is disabled, unavailable, or reports that work is not eligible, do not repeat the same action from another tab.

Retry supported failed jobs

Open Jobs > Failed Jobs to review the safe retry queue for server-owned media, reports, guide packages, native uploads, EHR exports, text-to-speech, and extraction work. Each row shows the subject, kind, status, failure category, retryability, attempts, last update, and available actions.

Choose Retry only when the row is marked retryable. CareMap may block retry when work is already running or complete, the source record is missing, the job has reached its attempt limit, or the previous partial result is not safe to resume. Use Audit when you need the event trail, and use the source link only if you have permission to open the related Report, Guide, Form, or Person.

The Native Uploads card summarizes background upload evidence jobs. The Recent Jobs card shows the latest server job activity so you can confirm whether a retry starts, succeeds, or fails again.

Escalate with safe details

When you need help, record the operation group, status label, failure category or reason, updated time, job kind, subject label, and attempt count. Do not share raw storage paths, signed URLs, notification tokens, provider payloads, credentials, or private care details through an unapproved channel.