Managed DBA¶
Service ownership
Owner: managed-services (managed-pm@clouddigit.ai) — Status: GA — Last audited: 2026-05-11
Database administration as a service. We tune your queries, manage your schemas, watch your alerts, and join your war rooms.
What it is¶
A standing DBA capability bolted onto Managed PostgreSQL, MySQL, MongoDB, Redis, Kafka, and OpenSearch. The platform-level service runs the engine; Managed DBA runs the workload on top.
Scope¶
| Day-1 (build) | Day-2 (run) |
|---|---|
| Schema design review | Slow-query analysis |
| Index advisory | Lock and contention investigation |
| Migration plan review | Vacuum / autovacuum tuning |
| Capacity sizing | Replication health monitoring |
| Backup-strategy design | Failover drills |
| Security baseline review | Patch staging |
| Performance baseline | Incident response (P1/P2) |
Engagement model¶
- Standing retainer with defined coverage (engines, regions, hours)
- Ticket-based intake; we open tickets against ourselves for proactive work
- Monthly health review report
- Quarterly capacity & performance review with your team
Pricing¶
Per engine-cluster-month, with bands by cluster size and on-call coverage (BWH-only vs 24/7). See Pricing.
Related¶
Operate this service¶
Cloud Digit DBA-on-tap: schema reviews, query tuning, capacity planning, and 24×7 DB oncall.
Engagement model¶
| Tier | Coverage | Response SLA |
|---|---|---|
dba-business-hours | 8 AM-6 PM weekdays | 4 h |
dba-extended | 6 AM-10 PM weekdays + weekday peak weekends | 1 h business / 4 h after-hours |
dba-247 | 24×7 | 30 min P1 / 4 h P3 |
What's covered¶
- Quarterly health reviews per DB instance
- Schema design consultations
- Query tuning sessions
- Production incident response (P1)
- Migration planning support
- Capacity planning
What's not covered¶
- Application code (we tune queries, not write apps)
- Long-term consulting projects (see Compliance Consulting for related)
- Database license selection / commercial negotiation
IAM¶
| Role | Can do |
|---|---|
dba.viewer | View engagement records |
dba.requester | Open consultation requests |
dba.admin | Manage engagement contract |
Onboarding¶
Initial assessment per database: 1. Schema review 2. Query workload profiling (24-h capture) 3. Backup/HA posture assessment 4. Recommendations report
Re-run quarterly.
Related¶
Engagement requests¶
bash cd managed dba request \ --type query-tuning \ --database acme-prod-pg \ --priority normal \ --description "checkout query degraded after 2026-05-01 deploy"
Categories: - health-review — periodic assessment - query-tuning — specific query optimization - schema-design — new schema review - incident — P1 production issue - migration — DMS / version-upgrade support - capacity-planning — sizing for upcoming load
Quarterly health review output¶
Per DB: - Throughput / latency trends - Top 10 slow queries (week) - Index utilization - Bloat / vacuum efficiency - Connection patterns - Backup / restore drill validation - Recommendations (with rough effort estimate)
Findings filed as tickets in your tracker.
Incident workflow¶
P1: Cloud Digit DBA pages within SLA. They: 1. Join customer war room 2. Diagnose, propose mitigations 3. Implement (with customer authorization) 4. Post-incident write-up
Co-managed — customer team owns the system, DBA assists.
Metrics tracked¶
| Metric | Notes |
|---|---|
dba.requests.open | Should be < 5 at any time |
dba.requests.mean_time_to_resolve | Track per category |
dba.health_reviews.completed_qtr | One per DB per quarter |
dba.recommendations.implemented_pct | Drives engagement value |
Knowledge transfer¶
Engagement notes form a per-DB knowledge base: - Recurring issues - Past tuning decisions - Open optimization opportunities
Visible to customer team + CD DBAs.
Related¶
SLA breach¶
CD missed the response SLA: - Auto-credit applied - Investigate cause; if pattern, escalate - Document in monthly review
Engagement value perception low¶
Customer feels they're not getting value: - Schedule a calibration session — review engagement requests, output, gaps - Adjust engagement scope (more/different requests) - Consider tier change (down or up) - Worst case: end engagement, switch to ad-hoc consulting
Request handed off / inconsistent quality¶
Different DBAs per request → inconsistent advice. Mitigate: - Designate a primary DBA per customer - Backup DBA familiar with the environment - Shared knowledge base for handoffs
Recommendation that customer disagrees with¶
DBA recommends X, customer believes Y: - Document both positions - Run a test if practical - Escalate to senior DBA if needed - Customer ultimately decides; DBA provides input
Disagreement is healthy when documented.
Incident-only customer (not using proactive)¶
Customers who only open requests during incidents: - Health-reviews backlogged - Recommendations not implemented - High incident rate ensues (predictable)
CE engages: schedule proactive engagement; the math favors it.
Onboarding stalled¶
Initial assessment incomplete (e.g., couldn't profile workload): - DB write permission for the DBA principal blocked profiling - DB at capacity; profiling load unacceptable - Org change; original sponsor left
Reschedule with the right access and authorization.
Custom DBA team requested¶
Some customers want named DBAs. Possible for high-tier engagements; talk to CE about retainer model.