Technology
Serverless and managed services for resident-facing applications
How agencies can run online services that scale for deadlines and cost little when idle, without managing servers.
Resident-facing services have uneven demand: a permit portal is quiet most of the year and busy before deadlines; a results site sees a year of traffic on one election night. Serverless and managed cloud services fit this pattern because they scale automatically and charge for use rather than for idle capacity.
What serverless means
The cloud provider runs the code, database, and messaging; the agency deploys application logic and configuration. There are servers, but nobody at the agency patches or sizes them. Managed databases, object storage, queues, and functions are the building blocks.
Benefits for agencies
- Automatic scaling for peaks without capacity planning
- Low cost for services with modest average traffic
- Reduced operational burden and patching
- Built-in high availability
- Faster delivery of new services
Trade-offs
- Architectural constraints; not every application fits
- Vendor-specific services can increase switching cost
- Cost can rise with very high sustained traffic
- Debugging and observability require appropriate tooling
- Cold start latency for infrequently used functions
Good fits
Online forms and applications, payment integrations, notification services, APIs for open data, scheduled data processing, and event-driven integrations between systems.
Poorer fits
Long-running batch jobs with heavy compute, applications that need persistent connections, and systems that must run on premises for regulatory reasons.
Security and compliance
Managed services inherit the provider's authorizations, but configuration remains the agency's responsibility. Identity, network access, logging, and data classification apply as in any architecture.
Cost governance
Budgets and alerts by service, since consumption pricing can surprise. Model peak scenarios before launch.
Where HQT fits
Architecture, implementation with infrastructure as code, integration with systems of record, and the observability and cost controls that keep a serverless service predictable.