Remote Hands (SLA-backed)¶
Service ownership
Owner: dc-operations (colo-pm@clouddigit.ai) — Status: GA — Last audited: 2026-05-11
SLA-backed remote operations — console attach, KVM-IP, port turn-up, configuration assist.
What it is¶
Operations work that's done on the network, not physically inside the rack. Faster than Smart Hands because there's no walk to the floor; backed by an SLA on response time.
Examples¶
- KVM-IP / IPMI session attach so you can run a recovery shell
- Switch port enable/disable
- Cable/port label verification (via reach to the in-rack monitoring)
- Loopback / link-light checks
- Coordinated network event support during your maintenance window
SLA¶
| Severity | Response |
|---|---|
| P1 (production down) | < 5 min |
| P2 (degraded) | < 15 min |
| P3 (request) | < 1 hour |
| P4 (scheduled) | Per schedule |
These map to the platform-wide Support tiers.
Pricing¶
Included as part of your colocation commitment up to a tier-defined monthly hour budget. Beyond the budget, billed at standard Remote Hands rate. See Pricing.
Related¶
Operate this service¶
Pre-purchased pool of DC technician hours with response-time SLAs — for customers with predictable smart-hands needs.
When this fits over Smart Hands T&M¶
- Monthly usage > 20 hours
- Predictable hardware refresh / maintenance cadence
- Need guaranteed response time (T&M is best-effort)
For occasional smart-hands work: T&M is more economical.
Plan tiers¶
| Plan | Monthly hours | Response SLA (P1) | Response SLA (P3) | Discount |
|---|---|---|---|---|
remote-50 | 50 h | 60 min | 4 h | ~15% |
remote-100 | 100 h | 30 min | 2 h | ~25% |
remote-200 | 200 h | 15 min | 1 h | ~35% |
remote-unlimited | unlimited | 10 min | 30 min | quote |
Unused hours partial roll-over (max 25% to next month); does not accumulate indefinitely.
IAM¶
Same as smart-hands; the plan applies to all smarthands.requester calls in the project.
Priority tiers¶
| Priority | Definition | Example |
|---|---|---|
| P1 | Production-down, customer-impacting | Failed drive on prod DB |
| P2 | Production at risk | Failing PSU still online |
| P3 | Routine work | Planned hardware upgrade |
| P4 | Documentation / non-urgent | DCIM cleanup |
Plan SLA applies per priority. Misuse of P1 for non-urgent burns goodwill.
Cost structure¶
Monthly fixed fee per plan. Overage hours (above plan limit) bill at T&M rate.
Related¶
Metrics¶
| Metric | Healthy | Alert |
|---|---|---|
remote-hands.hours_used_mtd | within plan | > 90% of plan |
remote-hands.sla.met_pct | > 99% | < 95% |
remote-hands.requests_open | low | growing backlog |
remote-hands.avg_response_time_min | within SLA | breach |
Request workflow¶
Identical to T&M smart-hands, but with SLA accounting:
```bash cd colo smart-hands request --priority P1 --rack
Response time tracked; SLA breach automatically credits to plan¶
```
Reporting¶
Monthly SLA report: - Hours used vs plan - Per-priority response times - SLA breach count + credit - Plan utilization trend
Helps decide upgrade/downgrade at next renewal.
Plan sizing¶
Track 6 months: - Average hours/month - Burstiness (max month vs average)
Plan capacity should cover average + 30% burst. Burst above that bills as overage; chronic overage means upgrade.
Coordination with project teams¶
Remote hands requests should follow ticketing: - Production incident → ITSM ticket → remote-hands request - Routine upgrade → planned-change ticket → remote-hands scheduled
Avoid bypass workflows; the audit trail matters.
Related¶
SLA breach¶
CD missed response time: - Auto-credit applied (hours back to plan, or BDT credit) - Investigate cause; if pattern, escalate to CE - Document in monthly review
Hours exhausted mid-month¶
hours_used_mtd > plan: - Continue using remote-hands; overage at T&M rate - Or downgrade priority (use external resource for low-priority) - Or upgrade plan immediately
P1 escalations¶
P1 not getting attention: - Verify request marked P1 (not P3 by typo) - Hotline call confirms severity - CE engagement for repeated misses
P1 is reserved for production-down; misuse erodes the SLA for everyone.
Underutilizing plan¶
Months of < 50% utilization → downgrade at renewal. Don't pay for capacity you don't use.
Caveat: utilization is bursty by nature; one quiet month doesn't mean downgrade. Look at 6-month trend.
Rollover not honored¶
Up to 25% of unused hours roll to next month: - Verify plan terms (cd colo remote-hands plan --show) - Some custom plans have different rollover terms - Disputes via CE
Pricing change at renewal¶
CD may adjust pricing annually: - Pre-notified 90 days before renewal - Negotiable based on usage history - Multi-year commitments stabilize pricing