Discovery
- Workload profiling across 18 customers
- Tenant isolation requirements
- Quota model design
SaaS · 2023
Observability that scales with the customer.
The challenge
CloudAxis was outgrowing single-tenant. Onboarding a customer took 14 days of manual work. Their architecture couldn't isolate noisy tenants from each other.
The approach
We re-architected the data plane for true multi-tenancy with per-tenant resource quotas, then rebuilt the onboarding flow as a self-serve product surface.
By the numbers
0 → 4 min
Onboarding time
0
New tenants in Q1
0%
Uptime
0x
Throughput
The process
From discovery to launch.
Architecture
- Cell-based data plane
- Per-tenant ClickHouse pools
- Quota enforcement at ingest
Design System
- Self-serve onboarding flow
- Dashboard builder with drag composition
- Density-aware data tables
Engineering
- Go ingest with backpressure
- Terraform per-tenant provisioning
- Feature flags for incremental rollout
Launch
- GA with 18 design partners
- Self-serve open for new signups
- Sales-engineering handoff complete
Onboarding flow · dashboard
From strategy to execution, they worked like an extension of our engineering org. We launched two major features ahead of schedule with measurable customer impact.
Elena RodriguezDirector of Engineering, CloudAxis
99%
Onboarding
220
New tenants
3.2x
ARR growth
Outcomes
The numbers settled. The story didn't.
Onboarding dropped from 14 days to 4 minutes. The platform absorbed 220 new tenants in a single quarter without operational strain. ARR grew 3.2x year-over-year.
Next case study