Six milliseconds across the Gulf.
Qafila runs dispatch and tracking software for freight operators in Saudi Arabia and the UAE. Their customers are on phones, in warehouses, and on bad connections.
Riyadh region, measured from Dubai and Dammam.
Versus the previous provider.
Over the last twelve months.
The problem
Qafila's app servers were in Frankfurt. Every dispatch screen paid a 90–120 ms round trip before it rendered, and mobile users on weak connections felt it. The previous provider's nearest region was still a continent away, and the bill included egress charges that grew with every new customer.
The move
We moved their API, worker queue and PostgreSQL primary to ruh1 in Riyadh. The database runs on a Scale instance with daily snapshots; two Growth instances handle the API behind a private network. Static assets are cached at the edge.
The migration ran as a read-replica cutover: replicate, verify, pause writes for 90 seconds, promote, switch DNS. The team rehearsed it twice in staging before touching production.
The result
Median API latency from Gulf networks fell to single-digit milliseconds. With egress included and no per-seat platform fee, the monthly bill dropped 41% at the same traffic level.
Omar K. · CTO, Qafila
Stack
Talk to the team.
We will walk through your current setup and give you a straight answer about fit.