Legal-services SaaS in Leeds
A case-management platform moves from a single PostgreSQL instance to streaming replication with automatic failover, after a disk failure once cost them an afternoon of downtime.
Managed PostgreSQL & ClickHouse Hosting
DevOps services · United Kingdom
Managed PostgreSQL hosting for UK teams who need a database that stays up, plus ClickHouse when the reporting queries start slowing down the transactional side. We run replication, point-in-time recovery backups and tuning on VPS and dedicated servers in Frankfurt, and we pair the two engines when your analytics no longer belong on the same tables as your checkout.
- PostgreSQL replication with tested failover
- Point-in-time recovery, restores actually rehearsed
- ClickHouse for analytics, kept off the OLTP path
- VPS 1–8 h/month; dedicated 5–10 h/month depending on plan; GPU 10 h/monthIncluded hours
- Frankfurt, in the EUData centre
- PITR backups with restores tested, not assumedBackups
- EUR invoicing, 20% UK VAT itemisedBilling
01
OLTP and OLAP, run as two different jobs
PostgreSQL handles your transactional workload; ClickHouse handles the reporting queries that would otherwise lock up your checkout tables. We set up change-data-capture from PostgreSQL into ClickHouse so your dashboards run on a separate engine built for scans across billions of rows, while the primary database stays fast for writes. Most teams only need this split once reporting queries start showing up in slow-query logs.
- Streaming replication with automatic failover on PostgreSQL
- PgBouncer connection pooling sized to your workload
- CDC pipeline from PostgreSQL into ClickHouse for analytics
- Partitioning and index tuning based on your real query patterns
02
Best practices we apply
Database incidents are expensive precisely because they're rare, so the checklist has to hold every time, not just on the day someone remembers to run it. We apply the same set on every PostgreSQL and ClickHouse deployment and hand you the runbook that proves it, so the next engineer — ours or yours — can verify it too. Restore drills are logged with a timestamp, not just claimed, so a questionnaire answer is backed by evidence rather than a promise.
- PITR restores tested on a schedule, not assumed to work
- Streaming replication with automatic failover (Patroni)
- PgBouncer pooling to stop connection exhaustion under load
- Least-privilege roles, TLS in transit, no public database port
- Autovacuum and index tuning reviewed against real query plans
03
UK compliance and data residency
Under UK GDPR and the Data Protection Act 2018, where your customer data sits and who can reach it both matter to your ICO obligations. We keep databases off public IPs entirely, reachable only through your application network or VPN, with encrypted connections and least-privilege access. Both data centres are inside the EU, which keeps the residency conversation simple for UK companies serving European customers.
04
How the included hours work
DevOps hours come with the plan: 1–8 hours a month on VPS tiers, 5–10 hours a month on dedicated servers depending on the plan, 10 hours a month on GPU servers. You request them from the client panel, describe the work — a replication set-up, a ClickHouse migration, an index review — and propose a deadline. Hours don't roll over, so teams typically use one month for migration work and the next for tuning. Work beyond the quota is accepted and billed separately, never silently dropped.
How we work
Request the work from the panel
You describe your current schema, your query patterns and your uptime requirements under DevOps Services, and propose a deadline.
We scope replication, backups and CDC
We confirm the replication topology, the backup and PITR window, and whether ClickHouse is needed now or can wait.
We migrate and rehearse failover
We set up the databases, migrate your data, and actually trigger a failover and a restore before calling it done.
We hand over with a tested runbook
You receive the topology diagram, the backup schedule, the failover procedure and the tuning notes from our first weeks live.
Use cases
E-commerce brand with growing order history
A retailer's reporting dashboards were slowing down checkout at peak hours; CDC into ClickHouse moved the analytics load off the transactional database entirely.
Public-sector supplier in Bristol
A contractor serving local authorities needs documented, tested backups and least-privilege database access to satisfy procurement due-diligence questionnaires.
Technology stack
- PostgreSQL 16
- ClickHouse
- Patroni
- PgBouncer
- Debezium
- pgBackRest
- Grafana
Every subscription includes monthly DevOps, CyberSecurity and Programming hours: 1 h from the cPanel Advanced plan up, 1–8 h on VPS, 5–10 h on dedicated servers depending on the plan, and 10 h on GPU servers. Hours are requested from your client panel, don't roll over to the next month, and anything above the quota is accepted and billed separately.
Frequently asked questions
Is managed PostgreSQL worth it compared to self-hosting?
It's worth it once database work starts competing with feature work — replication, PITR restores and tuning take real engineering time to get right and keep right. Managed hosting with included DevOps hours removes that ongoing burden without the usage-based pricing of the big cloud databases.
Can I run PostgreSQL and ClickHouse together?
Yes — this is exactly the split we set up. PostgreSQL stays your transactional source of truth; ClickHouse receives a continuous copy through change-data-capture for reporting and analytics, so heavy read queries never compete with your application's writes.
ClickHouse vs PostgreSQL — which do I actually need?
PostgreSQL for transactional data: orders, users, anything with frequent small writes. ClickHouse for analytics: dashboards, aggregations and scans across large historical datasets. Most growing applications end up needing both, connected by a CDC pipeline rather than a single database doing two jobs.
How are PostgreSQL backups actually verified?
We run point-in-time recovery backups and test the restore procedure on a schedule, not just at set-up. A backup nobody has restored isn't a backup you can rely on, so the restore itself becomes part of the checklist we document for you.
Is my database reachable from the public internet?
No. Databases sit off public IPs by default, reachable only through your application's private network or a VPN connection, with TLS enforced and least-privilege roles. This is one of the standard hardening steps applied on every PostgreSQL deployment.
Can you migrate our existing PostgreSQL database to you?
Yes, migration is free and handled white-glove, typically completed in under 24 hours with a rehearsed cutover to minimise downtime. We plan the migration window around your traffic patterns and confirm replication is healthy before switching over.
Want a database that survives a bad night?
Tell us your current setup and traffic pattern, and we'll design replication, backups and — if you need it — a ClickHouse pipeline, using the hours already in your plan.