Known Limitations
Overview
Section titled “Overview”Korta is production-deployed but still evolving. The current architecture prioritizes simplicity, low cost, and developer clarity over advanced scale features.
Current limitations
Section titled “Current limitations”Free-tier infrastructure constraints
Section titled “Free-tier infrastructure constraints”- Impact: Throughput, cold starts, and latency can vary under burst traffic.
- Workaround: Keep rate limits active, monitor usage, and avoid heavy background workloads.
- Planned fix: Move to paid resources and tune service sizing when sustained traffic grows.
Single-region backend/database setup
Section titled “Single-region backend/database setup”- Impact: No geo-redundancy or regional failover; outages affect all users.
- Workaround: Monitor health checks and keep operational runbook ready for fast recovery.
- Planned fix: Introduce multi-region strategy and failover planning.
Limited analytics depth
Section titled “Limited analytics depth”- Impact: Only basic click counts are available (no advanced dimensions like referrer/device/geo).
- Workaround: Use click totals for early-stage validation.
- Planned fix: Extend analytics schema and reporting endpoints.
No refresh-token/session invalidation layer
Section titled “No refresh-token/session invalidation layer”- Impact: JWT sessions are short-lived but there is no token revocation list.
- Workaround: Re-authenticate on expiration; rotate API keys when needed.
- Planned fix: Add refresh tokens and optional server-side session revocation.
Tight dependency on external services
Section titled “Tight dependency on external services”- Impact: OAuth, email delivery, and hosting provider issues can affect user flows.
- Workaround: Keep provider configs validated and maintain fallback operational steps.
- Planned fix: Add provider abstraction and stronger outage playbooks.
Manual operations and observability maturity
Section titled “Manual operations and observability maturity”- Impact: Incident response still depends on manual checks/log review.
- Workaround: Use
/api/v1/health, deploy logs, and structured checks after release. - Planned fix: Add automated alerting, dashboards, and CI/CD quality gates.