ServerBudget

What a Cheap VPS Does Not Include by Default: A Five-State Audit

By Marcus Reid · Published Nov 11, 2026 · Last updated Nov 11, 2026

Short answer: The cheapest tier does include root access, a dedicated IPv4, sixteen IPv6 addresses, a custom ISO, a provisioning API, VNC and rescue mode; those are printed in the all-plans table. What it does not publish is a default state for security and data safety. Backup is the one item with two conflicting published states - a plan page that presents server-level backup as protecting data integrity, and an acceptable-use policy that describes a weekly copy with no integrity guarantee, made for server restoration only. DDoS protection, managed firewall rules, OS patching, monitoring, malware scanning and customer-restorable snapshots are absent from the retrieved plan and policy pages altogether. An application's own official backup documentation says a usable copy needs five components, a maintenance-mode step and storage outside the installation. That gap is not a missing price. It is a missing responsibility, and on the cheap tier it lands on the operator.

ItemDefault state on the pages readWhere that state is publishedWho supplies it
Root access, 1 IPv4, 16 IPv6, custom ISO, VNC, rescue modeIncludedAll-plans tableVendor
Control panelPriced add-onHome page and plan page, two different figuresBuyer, at extra cost
BackupsPresent but limited, and described two different waysPlan page versus acceptable-use policyBuyer, for anything restorable
Key-only SSH, fail2ban, restricted firewallAbsent from the plan pagesOfficial setup guide for self-hosted softwareBuyer, by hand
DDoS, patching, monitoring, malware scanning, snapshotsNo published stateNowhere on the retrieved pagesUnknown from this record
Off-box copy, database dump, maintenance modeAbsent from the plan pagesOfficial application backup documentationBuyer, by hand

"Not published" is a finding here, not a euphemism for "not offered". These pages may describe such a service elsewhere. On the pages read, they do not.

1. Why two states are not enough

Most checklists sort features into included and extra. That binary hides the interesting cases. Five states appear in this record, and they behave differently at 2 a.m. when something breaks.

  1. Included - named in the all-plans table, no condition attached.
  2. Priced add-on - named and separately charged, so its absence is a budget decision.
  3. Present but limited - named with a qualification that changes what it can do. Backup is this state.
  4. Absent from the record - not named on any page retrieved. Could exist; cannot be assumed.
  5. Operator-supplied - defined by a third party's official documentation as work the person running the machine must do.

Only the first two are visible on a price page. A cheap tier looks complete if you stop there, because the table is genuinely long. It is the fourth and fifth states that decide whether a small project survives its first disk failure.

2. What the included table actually lists

The all-plans table on the plan page is unusually specific about virtualisation features. It lists a 1 Gbps network port, unmetered traffic with a fair-use footnote, one dedicated IPv4 address, sixteen IPv6 addresses, full root access, CPU and memory type descriptors, local NVMe storage, custom ISO support, resource metrics, status logs, configurable time zones, a provisioning API, always-on access, HTML and Java VNC, and rescue mode. [S1] A separate line prices a control-panel license as an add-on; the home page prints a different starting figure for the same category. [S1][S2]

That is a real list of capabilities. None of them is a security control, a patching schedule, a monitoring alert or a restorable copy of your data. Rescue mode is a recovery console, not a backup. VNC is access, not protection. Conflating "I can reach the machine" with "the machine can be restored" is the single most common misreading of a long feature table. For the security layer a base plan leaves unquantified, read what three VPS providers publish about DDoS protection.

3. The backup row has two published states

The plan page has a short section presenting server-level backup as a way to keep data intact and available from the control panel. [S1] The acceptable-use policy describes the same facility quite differently: changed data on VPSs is copied weekly, no guarantee of any kind is given as to the integrity of those copies, and the copies exist for server-restoration purposes only. It adds that if data loss is caused by a vendor error, recovery will be attempted at no charge. [S3]

Both statements are published by the same vendor. They are not contradictory in a legal sense - one is a marketing summary, one is a policy - but they are contradictory as an operational answer to "can I restore yesterday's database?" A weekly copy made so the vendor can rebuild a host is not a customer-restorable snapshot with a published retention period, and the policy explicitly declines to guarantee its integrity. The word "attempt" in the recovery sentence is a promise of effort, not a promise of a result.

Two more clauses bound the same item. The terms state that moving server content off the machines at termination is the customer's responsibility and that the vendor will not transfer or move that content elsewhere. [S4] The acceptable-use policy states that after a suspension lasting more than 14 calendar days, backup copies of the account are permanently deleted on termination, with no refund. [S3] So the copy is time-bounded by the account's standing, not by a retention schedule the buyer chooses.

A price comparison cannot see any of this. The add-on catalog in an earlier audit named backup products without prices; this article is asking a different question - not what a backup costs, but what state the default one is in.

4. What an application's own documentation requires

To make "restorable" concrete, read the backup chapter of a widely deployed self-hosted application rather than a hosting page. Its official administration manual says five things must be retained: the config folder, the custom apps folder if any are installed, the data folder, the theme folder, and the database. It describes enabling maintenance mode to lock sessions and prevent new logins during the copy, instructs that folders be copied to a location outside the installation, and gives database dump commands for MariaDB, MySQL, SQLite and PostgreSQL. [S5]

That is a five-component, two-step definition of a backup from the software publisher. Compare it with a weekly host-level copy with no integrity guarantee. The two are different objects. An operator who assumes the host copy contains a consistent database dump has no basis for that assumption in either document.

The same manual's existence also settles a subtler point: the operator is the intended actor. The procedure is written for the person administering the application, with commands they run. A hosting plan that includes a background copy does not execute those steps for them.

A related official setup guide lists key-only SSH, fail2ban and a firewall limited to SSH, HTTP and HTTPS as machine-preparation tasks. [S6] It publishes no hardware number, which is why it is used here as a task list rather than a sizing source. Those three items are security defaults that the cheap tier's feature table does not claim to perform.

5. What real operators built instead

Public discussion shows the same gap being filled by hand. These comments were retrieved through a public comment index, are linked below with author and date, and are self-reported anecdotes, not measurements or vendor statements.

One operator describes a home NAS pushing nightly encrypted snapshots to a rented VPS, with the VPS used as an off-site target rather than as the only copy. [HN1] Another, discussing a self-hosted file-sync server, warns that object storage used as a backing store must be configured before any users are added "otherwise you may lose data on disk", and reports using local storage with a separate copy to object storage. [HN2] A third runs nginx on a small VPS while keeping files on a separate storage service with a WebDAV interface, quoting an all-in monthly figure in euros for the combination. [HN3]

A different strategy appears in a thread about low-end machines used for business: instead of paying more for one reliable host, the author describes a "let it fail" approach with replication across many machines and a separate hardware firewall appliance. [HN4] And a 2011 comment names the attraction of VPS plainly - "guaranteed resources and the ability to debug and fix problems yourself" - while immediately warning about "undisclosed limits". [HN5] That sentence captures both halves of this article: capability is included, responsibility is transferred.

The consensus across these reports is that people build their own off-box copy and their own failure strategy. The divergence is architectural - off-site snapshots, object-storage backing, replicated cheap nodes - and none of the authors reports relying on a host-level copy they did not control.

6. Where the sources agree, and where they split

Agreement: the all-plans table is long on access and virtualisation features; the legal pages assign data-integrity and content-migration responsibility to the customer; an official application manual defines a backup as a multi-component procedure; community reports describe operators building that procedure themselves.

Divergence: the plan page and the acceptable-use policy describe the backup facility with different strength. The two published prices for the control-panel add-on differ. The terms and the acceptable-use policy both terminate backup copies on termination but phrase the trigger differently. The official application manual requires an off-installation copy; no hosting page retrieved states where its weekly copy lives or how a customer requests a restore.

Unresolved: no published retention period, no restore procedure for the customer, no statement on DDoS protection, patching, monitoring or malware scanning on the cheap tier, and no second or third vendor feature table readable today - seven pricing and product paths at one vendor returned HTTP 403 Managed Challenge, and the review roundup in the comparison set publishes no feature table. [S7][S8] An official forum thread that might have added a second application's backup procedure returned HTTP 404 on the path tried and is not quoted. [S9]

7. Who supplies the missing pieces

Read the included table as a statement about access, not about continuity. Then assign the remaining work explicitly: you provide the key-only SSH setup, the fail2ban configuration and the firewall rules; you provide the maintenance-mode step, the five components and the database dump; you provide an off-box destination and a restore test; you decide whether a weekly host copy with no integrity guarantee is worth counting at all. Items with no published state are not free and not absent - they are simply unverified, and an unverified control should not be in a recovery plan.

The useful question before buying a cheap tier is not "what does it cost?" but "which of these five states is each safety item in?" On the pages read, most of them are in state four or five. That is the default, and it is the part of the plan no price column shows.

Frequently Asked Questions

Does a cheap VPS include backups?

A host-level copy is described on the retrieved pages, but its published state is limited: weekly, without a guarantee of integrity, and made for server restoration rather than customer restore. A plan page describes it more strongly. Neither page publishes a retention period or a customer restore procedure, so it should not be treated as a restorable backup of your application.

Is DDoS protection standard on the cheapest tier?

No published state for it appears on the plan or policy pages retrieved for this article. That means it cannot be assumed, not that it is definitely absent. Other pages in the comparison set were unreachable, and a roundup publishes no feature table. Ask for the specific page that states the item before relying on it.

What do I actually have to configure myself?

The official setup guide for self-hosted software names key-only SSH, fail2ban and a firewall limited to SSH, HTTP and HTTPS. The official backup manual for a self-hosted application names five components to retain, a maintenance-mode step, a database dump and an off-installation destination. Those are operator tasks on the pages read, not items the cheap tier's feature table claims to perform.

Sources

All pages read on 2026-11-11 unless noted.

No sentence from the three compared sites is reproduced here; only published feature labels, figures and paraphrased policy facts are used. Community comments are quoted with attribution and are self-reported. No sentence here comes from a machine-generated transcript. No benchmark, traffic estimate, search-volume estimate, analytics figure or ranking judgement appears in this article, and the two data sources reserved for the 28-day report were not read.

Marcus Reid
Systems Architect & VPS Infrastructure Analyst

Marcus has been working with virtualized server infrastructure since 2015, deploying and benchmarking VPS and cloud instances across DigitalOcean, Vultr, Hetzner, OVHcloud, Contabo, RackNerd, Hostinger, and the major hyperscalers. His work focuses on the gap between what providers advertise and what actually happens when you deploy a real workload. He verifies every price and specification against official provider pages before publication.