Rescue
Performance incidents
Find the pressure point fast
When a query, workload or platform becomes a business risk, the work starts with evidence: execution plans, wait behaviour, bloat/vacuum, cutover risk and rollback safety.
Enterprise case studies · Auckland NZ · ANZ delivery
Proof that turns production pressure into delivery confidence.
A decision-ready proof room for AWS partners, hiring managers, CTOs and enterprise delivery teams. Each example shows the pressure, the intervention, the outcome and the practical signal behind Dipak Rijal’s database, cloud and migration capability.
28,800×PostgreSQL workload recovered from 30+ days to 1.5 minutes.
Zeroproduction downtime across 12 Oracle databases including Exadata.
AWShands-on support across RDS, Aurora PostgreSQL, DMS, DocumentDB and RDS Proxy.
70%performance gain in high-availability manufacturing and operations platforms.
What the case library proves
Fast signals for busy decision-makers: production ownership, cross-platform database depth and calm execution when the risk is real.
Modernise
Oracle · PostgreSQL · AWS
Move without breaking trust
Migration and upgrade work is planned around validation, rollback, stakeholder communication and production confidence — not just tooling.
Operate
Enterprise runbooks
Leave teams stronger after delivery
The best engagement produces more than a fix: runbooks, decision notes, knowledge transfer and repeatable patterns your team can keep using.
Production proof library
Each case is written for technical buyers and recruiters: situation, action, outcome and the engagement signal it gives.
PostgreSQL crisis rescue: from 30+ days to 1.5 minutes
A production PostgreSQL workload had moved from slow query to business incident. The priority was to recover performance without creating more risk.
Aurora PostgreSQLQuery tuningVacuum / bloatIncident recovery
ChallengeExtreme runtime was blocking business confidence and creating uncertainty for stakeholders.
ActionAnalysed execution path, data volume behaviour, bloat/vacuum pressure, safer plan options and rollback exposure.
ResultRuntime reduced to 1.5 minutes with a clearer path for ongoing stability.
Why it mattersShows production debugging depth under pressure, not theory-only database knowledge.
Kāinga Ora Oracle 12c → 19c upgrade across 12 databases
National-scale New Zealand Government database estate modernised from Oracle 12c to 19c, including Exadata, where production continuity was the non-negotiable requirement.
Oracle 19cExadataGoldenGateDR architecture
ChallengeCritical platforms needed modernisation without disrupting production users.
ActionPlanned replication, change sequencing, validation, cutover, rollback readiness and post-change monitoring.
Result12 databases upgraded with zero production downtime and sub-minute replication lag.
Why it mattersShows enterprise change discipline across legacy Oracle and modernisation pressure.
AWS database support: Aurora, RDS, DMS, DocumentDB and production operations
Inside AWS, the work required service-specific understanding, calm triage and clear communication for real enterprise workloads.
Amazon RDSAurora PostgreSQLDMSRDS ProxyDocumentDB
ChallengeEnterprise customers needed help across downtime, migrations, performance, configuration and reliability questions.
ActionDiagnosed service behaviour, wrote reusable runbooks, mentored engineers and guided technical stakeholders.
ResultBuilt production-focused depth across AWS database services and escalation-ready support patterns.
Why it mattersShows real AWS operating experience beyond certification-only cloud knowledge.
High-availability manufacturing systems: performance and stability uplift
Manufacturing and semiconductor environments need performance improvement without destabilising systems that must stay available.
SQL ServerOracle RACHA/DRRoot cause analysis
ChallengeHighly available systems required measurable improvement while protecting operational continuity.
ActionApplied root-cause analysis, performance engineering and structured production change execution.
ResultDelivered 70% performance gains in production-grade environments.
Why it mattersShows delivery maturity where downtime is expensive and trust matters.
AI-SDLC delivery engine: AI-assisted workflows and production documentation
Modern teams need speed without losing governance. The practical approach is AI-assisted analysis, specs, RCA, test planning and stakeholder-ready documentation.
AI-SDLCRunbooksAI-assisted RCADelivery workflow
ChallengeTeams need faster delivery while keeping traceability, review quality and production context.
ActionUse AI-assisted workflows for SQL tuning, RCA, specs, documentation and review loops.
ResultA repeatable delivery model for teams that need speed and confidence together.
Why it mattersShows future-focused capability that blends engineering depth with practical AI execution.
Built for the people who decide quickly
This page is not only a portfolio. It is a buying and hiring filter: it helps a partner, recruiter or technical leader understand where Dipak is useful before the first call.
For AWS / consulting partners
Senior subcontracting depth for database migrations, performance rescue, HA/DR design and delivery assurance when a project needs specialist confidence.
For recruiters / hiring teams
Clear evidence of ownership, enterprise pressure handling, legacy-to-cloud depth and stakeholder-ready communication.
For CTOs / platform leads
A practical way to evaluate whether the problem needs tuning, migration design, operational stabilisation or a short senior advisory engagement.
For delivery managers
Work is framed around runbooks, risk control, sequencing, validation and handover — so the team can keep using the outcome after the engagement ends.
The delivery operating model
A simple production-first pattern behind the case studies.
01
Stabilise
Protect production, capture symptoms and reduce operational risk first.
02
Diagnose
Trace the issue across database, cloud service, workload, query and application behaviour.
03
Design
Create a practical remediation or migration path with rollback, validation and stakeholder clarity.
04
Execute
Implement with runbooks, change control, monitoring and communication.
05
Handover
Convert the fix into reusable patterns, notes and operating discipline your team can keep using.
Need applied help on a real database, cloud or delivery problem?
Share the platform, the risk, the current blocker and what “success” looks like. I can help shape the first-step plan: performance triage, Oracle → PostgreSQL migration, AWS database architecture, HA/DR validation, cutover readiness or production runbook cleanup.
Assess riskPlan the pathDeliver safelyHandover cleanly