Payments are experiencing issues due to temporary restrictions in Russia. If your payment does not go through, please submit a support request.Our support team is available 24/7 — we are always here to help with hosting and server issues.We are now accepting requests for dedicated server rental and colocation services in our data center.Reminder: we recommend enabling backups for additional data protection.A new VPS/VDS lineup with NVMe storage and improved performance is now available.Maintenance work on some servers has been completed. All services are operating normally.
Article7 min readViews0

Administration is included in the cost—who fixes the online store?

The control panel, updates by request, and round-the-clock support leave various tasks for the website owner. A hypothetical failure helps test the limits of VPS maintenance even before choosing a service.

Comments 0

Open server on a blue service mat, with the removed cover and screwdriver nearby.
In this article

The VPS offer states that administration is included in the rental cost, yet when an order fails, support asks you to contact the developer. Is this necessarily poor support? No: server maintenance and store repairs may be separate services. However, discovering this boundary in the middle of a failure is a rather inconvenient way to read the tariff terms.

When choosing, I would check not the word 'managed', but the path for a specific task: who will notice the problem, who has the right to make changes, and who will hand over the work if the cause lies outside their authority. A useful outcome of such a conversation is several agreed-upon actions with assigned responsibilities. It is precisely these that show which part of operations you are actually handing over to the provider.

The panel is installed. This does not assign a duty administrator.

A ready-made image with a control panel reduces initial setup. Through the panel, you can create sites and perform predefined operations, but the presence of buttons does not mean someone else will monitor updates, troubleshoot errors, or verify recovery. The tool and the support service should be evaluated separately, even when sold as a single bundle.

This distinction is explicitly confirmed, for example, in the PS Cloud Services documentation: the self-administration model applies to VPS instances based on ready-made templates with a preinstalled control panel. Work inside the virtual machine remains the client's responsibility. This is a condition of a specific service, not a statement about all hosting with a control panel.

With another provider, part of the assistance may be included in the rental. In the ishosting documentation, standard administration, including OS updates and backup configuration, is described as on-request work. The same document lists restrictions related to non-standard changes to the stack and control panels. Both examples were verified on September 27, 2026. Their meaning is not to declare a winner: an equally attractive term in a tariff card does not guarantee the same level of service.

Read the 'on request' terms especially carefully. This is a normal service model if the store has someone who knows when and what to request. If you expected to also hand over initiative—problem detection, update planning, result control—you must agree to that separately. Otherwise, technically available support may never activate at the right moment.

A store has three distinct layers of work

The first layer is infrastructure: the physical node, network, and virtualization platform. The second is the environment inside the VPS: operating system, web server, PHP processing, and database. The third is the store itself: code, modules, cart rules, and integrations. This division is useful for discussion but does not define a universal contract boundary: one provider may handle multiple layers, while individual components may reside with other suppliers.

Consider a training scenario. After changing the server environment, the catalog opens but orders are not saved. The available virtual machine confirms only partial functionality. An administrator can determine that the request reached the application and ended with an error; a developer can identify a module incompatibility. Meanwhile, the store owner must decide whether it is acceptable to temporarily restrict purchases if restoration requires a shutdown.

In this scenario, you cannot preassign blame to an update, a module, or the hosting. Observations are needed: what changed, when the error appeared, and which operation reproduces it. A clear boundary is required to organize the investigation, not to shuttle the store owner between three different contacts.

A good support condition describes the task handover. For example, an administrator records the time and result of diagnostics, identifies the affected component, and passes the details to the assigned developer. The developer confirms receipt and the next step. Before confirmation, it must be clear who continues to handle the ticket. The phrase 'our part is working' leaves the owner with too much work if no one explains where the next part begins.

Three diagnostic zones: infrastructure, VPS environment, and the store application; during handover, preserve the symptom, the test result, and the next responsible party.
Training example of diagnostic handover between zones. Arrows show one possible route, not a mandatory sequence. The scope of work and responsible parties are agreed upon for a specific store.

Analyze one change before purchasing the service

Consider updating the PHP version in a store on 1C-Bitrix: Site Management. Installing the package and ensuring the service starts is only part of the possible work. You need to know the compatibility of the installed edition, modules, and customizations, choose the timing for the change, and verify the actions affected for shoppers. This does not propose performing the update; it is an example for checking the scope of support.

Ask the executor to describe how they will carry out this task specifically in your configuration. Who gathers compatibility information? Who prepares the test environment? Who checks the order after the change? If another team supports the code, how do they participate and who approves the overall plan? The answer "updates are included" does not replace answers to these questions.

Separately clarify the rollback to the previous state. Downgrading one component version does not necessarily restore the entire system: settings, dependencies, or data may have changed during the work. The recovery plan must correspond to the actual changes. Who prepares it, who verifies the availability of the necessary copies, and who has the authority to decide on the rollback are also part of the service, not minor organizational details.

For this scenario, a short list of questions is sufficient:

  • Who initiates the work: the provider based on their own observation, or only after a store request?

  • Which compatibility checks, test environments, and backups are included in the agreed scope?

  • Which actions require separate approval, payment, or developer involvement?

  • Who verifies the result on the store side, and to whom is the task transferred in case of an application error?

Written responses must refer to the named versions, modules, and access levels of your project. A list from a promotional page is useful as a starting point for discussion. Confirmation for a different operating system or a standard site without your customizations does not resolve the same question.

Support via ticket and store monitoring

Round-the-clock ticket intake answers the question of whether you can contact support at night. It does not, by itself, confirm continuous monitoring of the checkout process. Similarly, configuring availability checks does not mean a specialist will initiate recovery without a ticket. You need to determine what signals the executor receives and what action each signal triggers.

You do not need to purchase the same depth of control for all parts of the site. For a small project, coordinated infrastructure monitoring and a clear ticketing procedure may be sufficient. For a store with orders running around the clock, the absence of a person between a signal and an action can become a significant limitation. The selection criterion is your response process, not the number of monitoring badges in the description.

In this hypothetical example, the catalog opening check might still pass. Therefore, it is useful in the proposal to specify the exact result that matters to the business: for instance, the availability of an approved, secure checkout scenario. Its configuration, limitations, and signal handling are discussed separately. Monitoring must not create real orders, deduct funds, or send test messages to buyers.

Scope of work may change along with the site

The store connected a non-standard module, moved the database to a separate service, or changed panel settings. The previous maintenance description may no longer match the actual system. Agreeing on changes is important not only for security: the executor must understand which configuration they are obligated to maintain.

Ask in advance which modifications fall outside the supported set and what happens after they appear. Separate work estimates, transferring part of the tasks to the developer, or revising the service are possible. None of these options makes the proposal bad on its own. The problem arises when the store considers a component to be under maintenance, but the executor learns about it only from an emergency call.

For the next conversation with the provider, take one real type of change and one possible failure that follows it. If the response makes it clear who prepares the work, who acts when an error occurs, and who confirms the restoration of the required function, the term "administration" has acquired practical meaning. If only a support address is known at this point, the distribution of work still needs to be agreed upon.

Discussion 0

Share your experience and ask questions. Comments without links appear after editorial review.

No comments yet. Start the discussion.