TWZNBooking Q4 2026
Practice03 / 04
03Infrastructure

Keel

The infrastructure layer your business stands on.

The backbone your business runs on.

  • Queues & background work
  • Fallbacks & idempotency
  • Observability
01Thesis

Built to Take the Hit.

Nobody notices a backend until it fails. Keel is how we make sure they never do — or more honestly, that when a dependency fails, nobody can tell. We engineer for the bad day: caches that fail soft, payments that can never strand money in limbo, files that move in gigabytes without touching server memory, and security enforced at the database layer where compromised application code can't reach it. Boring reliability, achieved deliberately.

02Capability
01 / 04

Fail Soft, Never Loud

On Intouch, Redis going down is a non-event. Every cache operation is wrapped in graceful degradation: on a miss or an outage the app falls through to the database — slower, never broken. Tiered TTLs tune each domain; pattern-based invalidation keeps data fresh; rate limiting bypasses rather than blocks when the cache is unreachable. Availability is a design decision, made everywhere.

Fig.Graceful cache degradation
HitDegraded
Request
api route
Cache layer
Redis check
availability verified first
Cache hit
returned instantly
60s orders30m products24h reference
Miss or down
Postgres
slower, never an error
Response
always 200
6 events · play, or hover a node to trace its path
Users experience slower responses, never errors — availability over strict enforcement.
Graceful cache degradation — IntouchView case study
03Capability
02 / 04

Money Never Lands in Limbo

Payment systems earn trust in their failure paths. On PM Competitions, tickets are pre-allocated with row-level locking, payment confirms them, and any failure — a declined card, a dropped connection — rolls the allocation back automatically. On EVÖQ, every Stripe webhook is logged with a unique event ID: duplicates are ignored, failures preserved and replayable. No duplicate tickets, no phantom orders, no ambiguous states.

Fig.Payment with rollback
SuccessFailure
Step 1
Pre-allocate
reserve tickets, order pending
Step 2
Payment intent
card charged once
Confirm
order complete, tickets live
Rollback
order cancelled, tickets released
Pool consistent
none sold twice, none lost
6 events · play, or hover a node to trace its path
Tickets are reserved before the charge and released if it fails — the pool can't drift.
Pre-allocate → pay → confirm/rollback — PM CompetitionsView case study
04Capability
03 / 04

Heavy Lifting Without Heavy Infrastructure

LOCKD moves 5GB files with zero job queues, zero message brokers, zero persistent workers. Uploads pick one of three tiers — presigned, chunked multipart, streaming — so file bytes never burden the API server. ZIP archives stream from bucket to bucket through Lambda with no intermediate buffering, reporting back by webhook. Stale jobs are detected lazily on the read path. The developer experience of a job queue, with none of the infrastructure.

Fig.Upload tiers
Direct to S3Fallback
File picker
drag & drop · any size
At selection
Size check
routes each file by size
Presigned PUT
< 50 MB · one request
Chunked multipart
50 MB – 5 GB · 10 MB parts, resumable
Streaming fallback
server-proxied for edge cases
Private bucket
S3
files land encrypted, never public
6 events · play, or hover a node to trace its path
The upload strategy is selected automatically from the file's size — no user intervention required.
Three-tier upload strategy — LOCKDView case study
05Capability
04 / 04

Security Below the Application

We put security where application bugs can't undo it. Row-level security on every table, so even a compromised API can't leak data across users. Draw logic in stored procedures, out of reach of the application layer. PIN login hardened against timing attacks — every comparison runs in parallel, so response time reveals nothing. Files identified by magic bytes, not by the extension an attacker chose.

These are the rails Forge

Fig.Transfer lifecycle
GateLimitPurge
Recipient
opens the share link
Password
optional passphrase
Expiry timer
24h · 7d · 30d
Download limit
max per recipient
Download
short-lived presigned URL
Auto-cleanup
files purged from S3
Final state
Expired transfer
files gone, audit trail kept
6 events · play, or hover a node to trace its path
Every control is enforced at the API level — the frontend only reflects the current state.
Transactional access control — LOCKDView case study
06Also in scope

After launch

  • 01Ongoing Support

    We stay on the bridge.

    Launch is the start of the relationship, not the end. Keel engagements continue as long-term stewardship — monitoring, evolving, and hardening systems as the business grows. The systems above are all under active care today.

07By the numbers
  • job queues, 0 brokers, 0 workers
  • single-file uploads
  • on every table
  • payment states, fully audited
09Engagement
Options04

Ways to start

Every engagement starts with a conversation. Fixed-price entry points, then work scoped to the system it has to carry.
  • 01Retainer

    Watch

    £500/ month

    Monitoring, dependency updates, security patching, and a response SLA. The system stays healthy; you stay informed.

  • 02Retainer
    Most engagements

    Steward

    £1,000/ month

    Everything in Watch, plus an engineering day each month — small features, performance work, hardening. The system doesn't just survive, it improves.

  • 03Retainer

    Partner

    from £2,000/ month

    Embedded ongoing engineering: roadmap ownership, priority response, and capacity that scales with the business. We stay on the bridge.

Build work
Backends & Infrastructure
Scoped per project

Backends, infrastructure, and the reliability engineering above are priced against their own scope — every build is quoted against the system it has to carry.

Scope a build

Not sure where you fit? Tell us about the system and we'll point you at the right start.

Talk to us
10Other practices
Practices04
11Contact
London--:--:--

Need a backend that takes the hit?

If downtime, data integrity, or scale is what keeps you up at night, that's our home ground. Tell us where it hurts.

  • Backend Systems
  • Infrastructure
  • Reliability
  • Support

What happens next

  1. 01ReplyWe reply within 48 hours
  2. 02CallA 30-minute call to understand the problem
  3. 03ProposalA written proposal with scope and price