Tools
9
Monthly tooling cost
Not shared
- Google Cloud Storage Storage
- Google Cloud Run Hosting
- Neon Database
- Stripe Payments
-
FlightStack Developer tools - Firebase Cloud Messaging Other
- Google Analytics Analytics
- GitHub Developer tools
- Redis Database
Stack analysis
Generated by Gemini on Nov 14, 2026. Treat it as a starting point, not advice.
FlightStack uses a serverless Google Cloud architecture alongside Neon and Redis to coordinate self-hosted Flutter build runners. The stack is cost-effective at small scale, but data egress costs on build artifacts and database connection spikes during concurrent runner executions represent critical operational risks.
What could break
- Artifact storage egress cost spikes Medium Google Cloud Storage charges standard internet egress fees whenever build runners or end users download large iOS IPAs and Android APKs. Sustained build activity can lead to unpredictable storage bandwidth bills.
- Connection exhaustion between Cloud Run and Neon High Rapid scaling of Google Cloud Run instances during bursty webhook periods can open numerous concurrent connections to Neon. Without strict pooling, this can exhaust database limits and drop build status updates.
- Volatile agent queues on unmanaged Redis High If Redis is operated as a single non-replicated instance without persistent snapshots, a crash or restart drops pending build jobs and runner heartbeat state, causing orphaned builds.
- Firewall blocking of push notifications on self-hosted agents Medium Relying on Firebase Cloud Messaging for agent signaling often fails inside enterprise private networks or corporate VPNs that restrict outbound ports for Google push services.
Ways to optimize
- Migrate artifact storage to Cloudflare R2 Switching build artifact storage from Google Cloud Storage to Cloudflare R2 removes egress fees, significantly lowering costs as build size and download frequency increase.
- Enable Neon connection pooling Route all Cloud Run database traffic through Neon pooled connection strings to prevent serverless instance concurrency from exceeding database connection quotas.
- Use managed serverless Redis Replace self-hosted or static Redis with a managed provider like Upstash to handle variable connection loads gracefully alongside Cloud Run without idle server maintenance.
- Replace Google Analytics with privacy-focused alternatives Developer-focused audiences frequently block Google Analytics via browser extensions. Switching to a lightweight alternative improves metrics accuracy and avoids cookie consent banners.