A healthcare platform's cloud bill, cut by almost half.
Sehat Kahani, a telehealth platform in Pakistan. I ran its Azure infrastructure as DevOps and data engineer.
- RoleDevOps and data engineer
- WhenApr 2024 to [TBD]
- StackAzure, Kubernetes [TBD confirm]
The problem
The bill had grown faster than the product. [TBD: what was driving the cost, why nobody had fixed it, and what was at stake for a healthcare platform.]
What I changed
Most of the saving came from two boring things: scaling to real traffic and removing what nobody used. [TBD: order and detail.]
The result
Monthly cloud spend fell 45%, while production kept serving high traffic. [TBD: baseline, period, and what held steady: uptime, latency, compliance.]
The changes, in order
- Scaling that matched real trafficSized compute to what the platform used, not what it was provisioned for. [TBD details][TBD]
- Removed what nobody usedFound and retired underused resources across environments. [TBD details][TBD]
- [TBD] Third changeThe one that needed a human to sign off.[TBD]