Backup, snapshots and disaster recovery

Daily snapshots of the whole server, geo-redundant backups, RPO 1h, RTO < 1h. All included in hosting — no "premium backup" surcharge.

A backup you actually use

Most companies find out their backup doesn't work only when disaster strikes. Our philosophy: automatic snapshots, regularly tested on staging, restored with one click. No tapes, no complex tooling. Every week a random backup is restored to a staging server and compared against the original — you get a report by email.

Snapshots — frequency and retention

A snapshot is a full copy of the VPS or server state at a moment in time. We can run them more often for premium plans. For databases we add WAL streaming every 1-5 minutes — RPO drops to < 30 seconds.

Plan	Frequency	Retention	Location
cPanel Hosting	Every 24h	7 days	EU (Frankfurt)
VPS standard	Every 24h	7 days	EU (Frankfurt)
VPS managed	Every 6h	14 days	EU multi-region
Dedicated server	Every 6h + WAL every 1h	30 days	EU multi-region + external S3

RPO and RTO — hard numbers instead of promises

Instead of "backup ASAP" — concrete values. RPO (Recovery Point Objective) tells you how much data you can lose. RTO (Recovery Time Objective) tells you how long restore takes.

RPO

≤ 1 hour

Maximum data loss on failure — 1 hour of work. With WAL log streaming RPO drops to < 30 seconds.

RTO

< 1 hour

Full server restore from snapshot typically takes 15-45 minutes depending on size.

MTBF

> 99.99%

Mean time between failures. Our DC has dual power, redundant networking and ECC RAM.

Disaster scenarios — and how we handle them

Disasters are rare but high-impact. Here are the most common scenarios we cover, with concrete recovery times. Every VPS and dedicated server is covered by these procedures with no extra fee.

Hardware failure (disk, RAM)

The most common cause of data loss. Our DC uses RAID 10 + ECC RAM, so physical failures are detected before data loss. Hot-spare replica activates automatically.

RTO: 5-15 min

Ransomware or account compromise

Attacker encrypts server files. The pre-attack snapshot is immutable — stored in a second DC, isolated from the production account.

RTO: 30-60 min

Accidental data deletion

"DROP TABLE users" run against production. The most common human error. Point-in-time recovery (PITR) lets you go back to exactly 10 seconds before the mistake.

RTO: 15-30 min

Full DC outage

Fire, flood, extended operator outage. Managed plans have a replica in a second EU DC. Automatic failover on primary DC outage.

RTO: 30-90 min

App bug corrupts data

A deploy introduces a bug that corrupts DB records. Restore only the DB (not the whole server) — app files stay, data rolls back.

RTO: 20-45 min

Migration to another server

You outgrow the current plan and want to move to a bigger one. Snapshot → restore on the target VPS. Zero data loss, short downtime.

RTO: 15-30 min

Database backup — per-engine strategies

A full VPS backup is not always enough. Databases have their own point-in-time recovery (PITR) mechanisms — they let you go back to exactly the chosen moment, not just the last snapshot. We configure these free on managed plans.

PostgreSQL

WAL streaming + base backup every 24h. PITR up to 30 days back.

RPO < 30 sec

MariaDB / MySQL

binlog streaming + xtrabackup every 24h. PITR up to 14 days back.

RPO < 1 min

MongoDB

Oplog streaming + mongodump every 24h. PITR via oplog replay.

RPO < 1 min

Redis

AOF (Append-Only File) + RDB snapshot every 6h. Loss < 1 sec with fsync=always.

RPO < 1 sec

ClickHouse

BACKUP TO disk + replication. Snapshot per partition.

RPO < 5 min

Backup included in hosting — no hidden costs

We benchmark against the standard offering — most competitors charge extra for backup. With us everything is included, including geo-redundancy on managed plans.

Snapshot every 24h + 7-day retention — free on every plan
Snapshot every 6h + 14-day retention — free on managed plans (VPS 32G+)
WAL streaming for databases — free on managed and dedicated plans
Geo-redundancy (second EU DC) — free on dedicated plans
Snapshot export to external S3 (AWS, MinIO) — 0.05 PLN/GB
Extended retention (60/90/365 days) — 0.03 PLN/GB/month

Restoring from snapshot — in 4 steps

Open the panel

Pick a point in time

The list shows date, time and size of each snapshot. Pick the one before the problem.

Click "Restore"

You can restore to the same VPS (overwrites current state) or a new VPS (for comparison).

Done in 15-45 min

The server returns to its pre-incident state. Email confirms the restore.

Frequently asked questions

Does a snapshot require stopping the server?

No. Snapshots run online without interrupting the VPS. The app keeps running. For database consistency we use filesystem snapshots (LVM/ZFS) or table freeze in MariaDB/Postgres.

How long does restoring a very large snapshot (e.g. 500 GB) take?

On NVMe SSD: 30-60 minutes. From external S3: 60-120 minutes (bandwidth-dependent). We can also do "live restore" — VPS starts immediately, data is pulled in the background prioritizing actively-read blocks.

Are backups encrypted?

Yes. AES-256 at-rest encryption. Keys stored in Vault, access only for authorized restore operations. On enterprise plans you can BYOK (bring your own key).

What if I want retention longer than 30 days?

Up to 365 days possible via the customer panel. Cost: 0.03 PLN/GB/month extra retention. For regulated industries (banking, healthcare) we configure 5-7 year retention.

Can I download a snapshot locally?

Yes. Export a snapshot as tar.gz or via rsync. We can also push directly to your AWS S3 / Google Cloud Storage / Azure Blob.

Your data is too valuable to leave to luck

All 4hosting plans include backup. For dedicated and managed plans we add geo-redundancy and WAL streaming. See the pricing for details.

See plans →