What it takes to keep a game online.

A live game's backend has to stay up through server failures, launch days and the minute every player opens an event. Here is what that involves, in plain words, and how Lonca handles it, whether you use Lonca Cloud or host the open-source core yourself.

Six jobs behind a backend that stays up

High availability

The backend keeps answering even when one of its servers goes down.

What it means: Every part of the backend runs in more than one copy, in more than one place, so losing one copy doesn't take the game down.
What players notice without it: Players can't sign in, saves don't load and a running event disappears, usually at the busiest moment of the week.
Doing it yourself: Two or more copies of every service, load balancers, health checks and a database that fails over on its own, paid for around the clock.
With Lonca: Every part of Lonca is designed to run in more than one copy. On Lonca Cloud the target is 99.9% availability a month, with one platform team on call for every game.

Fault tolerance

When something breaks, nothing a player earned is lost or handed out twice.

What it means: Servers restart, networks drop and requests get retried. A fault-tolerant backend either finishes the work in progress or cleanly undoes it.
What players notice without it: A crash mid-event wipes progress, a retried request grants a reward twice, and the support inbox fills up.
Doing it yourself: Retries that can't double-apply, transactional event delivery and recovery runbooks, written and tested for every feature.
With Lonca: A save or a reward is confirmed only after it is safely stored, rewards are claimed exactly once, and caches rebuild themselves from durable logs. The recovery target is under 15 minutes.

Horizontal scaling

Adding more servers when more players arrive, instead of buying a bigger one.

What it means: One server has a ceiling. A horizontally scalable backend spreads players across many identical servers and adds more as traffic grows.
What players notice without it: A store feature or a successful ad campaign turns into slow loading screens and timeouts, right when new players decide whether to stay.
Doing it yourself: Servers that keep no player state, data split so no single row or machine becomes a bottleneck, scaling rules, and load tests before every big launch.
With Lonca: Lonca's API servers keep no player state, so more can be added at any time, and the data layout is designed for a game at 1M daily players.

Traffic spikes

The minute an event starts, everyone opens the game at once.

What it means: Live games don't get steady traffic. Event starts, broadcasts and daily resets send a wave of requests within seconds.
What players notice without it: The wave hits the same few records: one leaderboard, one team score, one event. The database locks up, and the event meant to bring players back pushes them away.
Doing it yourself: In-memory leaderboards, batched writes, queues and rate limits, tuned for every feature that can spike.
With Lonca: Busy writes such as scores and team totals go to memory first and reach the database in batches, and rate limits per player and per game keep one busy game from slowing the rest.

Security and isolation

Each game's players and data stay separate and protected.

What it means: A backend holds accounts and personal data. It has to keep attackers out and keep every game's data apart.
What players notice without it: A leak or a hijacked account costs player trust, and can break privacy law.
Doing it yourself: Password hashing, signed short-lived tokens, encrypted connections, access checks on every request, and tooling for privacy requests.
With Lonca: Passwords are hashed with Argon2id, access tokens are signed and expire after 15 minutes, every connection uses TLS, and every query is scoped to one game. Export and deletion requests run across every system.

Running it every day

Monitoring, alerts, upgrades and someone awake at 3 a.m.

What it means: A backend is never finished. It needs monitoring, alerts, security updates, database care and upgrades.
What players notice without it: In a small studio, one backend engineer becomes the on-call team, the database admin and the release manager, and features wait.
Doing it yourself: Dashboards, alerting, an on-call rotation, upgrade planning and capacity reviews: at least two engineers to cover nights and holidays.
With Lonca: On Lonca Cloud, one platform team runs it for every studio, so you get the backend without the pager. Self-hosting? You run it, on the same design, with usage metering built in.

What Lonca is designed to hold

These are design targets for Lonca Cloud, not measured results yet. Self-hosted, they depend on your infrastructure.

Availability: 99.9% a month
Confirmed saves and reward grants lost: None
Recovery after a failure: Under 15 minutes
Reads, server-side, 99th percentile: Under 50 ms
Score submits, server-side, 99th percentile: Under 80 ms
Sized for: A game at 1M daily players, about 100k online at peak

Why it costs less when games share

Doing all of this for one game means paying for spare servers that sit idle most of the day. Lonca does it once for many games, so the spare capacity is shared.

Today: a stack per game

Game A

Game B

Game C

Lonca: games share pods

Three games, three shared pods instead of nine, each one fuller. Every game's data, rate limits and keys stay isolated. Illustration, not a measurement.

See what it would cost your game

Worried about your next launch?

Tell us your daily players and your busiest moments, and we'll walk through how Lonca would carry them.

SDKs for Unity, Godot and web games.
Open source: Lonca Cloud or self-hosted.
Built for casual and mobile games with live ops on day one.
A founder answers every signup.