Your environment
An environment is one company's isolated place to run its applications. It has its own addresses, its own database, its own files and its own backups. Open it from Hosting → Environments.
Choosing your hosting size
Your project deserves enough room for its applications and the databases and supporting services that keep them running. We recommend Professional (2×) as the starting point for most projects.
- Essentials · 1×: for smaller projects and lighter workloads. Tighter allocations leave less room for demanding applications or growth.
- Professional · 2× — recommended for most projects: a balanced starting point, with more room for applications, databases and supporting services.
- Premium · 3×: extra headroom for larger workloads, more applications and growth, with 50% more capacity than Professional.
| Size | Capacity in your company’s pool |
|---|---|
| Essentials · 1× | 4 vCPU, 4 GiB memory, 40 GiB NVMe |
| Professional · 2× | 8 vCPU, 8 GiB memory, 80 GiB NVMe |
| Premium · 3× | 12 vCPU, 12 GiB memory, 120 GiB NVMe |
These names describe one, two or three standard units of capacity, with the same managed hosting foundation at every size. Premium buys extra capacity, rather than priority support or faster individual CPUs. See the current hosting prices before choosing, or ask us to help size your project.
One environment can run multiple applications
Purchased resources form your company’s pool. You can split a larger pool between environments when each environment’s minimum resource requirements fit; splitting the pool does not create more capacity. A standard 1× allocation typically uses the available capacity for one environment. That environment can run multiple applications, each with its own files, address and database, within its resources. There is no fixed site count for these standard sizes: what fits depends on the applications’ requirements.
Applications, database engines and supporting services all draw from the allocated resources. Premium’s 12 vCPU is the pool total, with an 8 vCPU ceiling per environment. More capacity gives you more room to allocate; it is not a guarantee of a particular visitor count or application performance.
Overview
Open Overview to see the selected environment by name. The health line shows whether a change is still applying. Below it, open a running site or its administration controls from Applications. An empty environment offers Set up a website, which opens application setup in that environment.
Switch to Tools for shared browser tools, or open Troubleshooting and technical details when you need to investigate a problem.

When a change is still being applied
Changes are applied in the background. A change is not lost if it takes a moment:
- Applying — the change is queued or in progress. Keep working; the page updates.
- Failed — the change could not be applied. The failure is recorded, and the environment may have applied only part of the change. Check its current state and contact support if service is affected.
- Applied — the page reads "Ready: every change is applied."
If a change fails, choose Apply again and read the message it reports. It applies the same recorded settings again, so you do not enter anything again; nothing is recreated and no site is reinstalled.
If two people change the same environment at once, one of them is told the change is stale rather than overwriting the other. Reload and make the change again against the current state.
Site insights
Open your environment from Hosting → Environments, then choose Overview → Insights. Each card covers one application's main custom domain. If Hosting asks for a TXT record, add the displayed name and value at your DNS provider. Select Check the record now to recheck after publishing, or wait for Hosting to check automatically.
Once the domain is verified, the card starts with one sentence about the site and the first thing to do, with a button that takes you to it. Monitoring and audit details holds the uptime, certificate and audit counts.
Fix a finding
Each finding says what is wrong and how to fix it. Where Hosting owns the fix, select Fix it. Where the fix is code, open Show the HTML to add (or the robots.txt or redirect equivalent) and select Copy code. After you change the page, select Check this page again. The card says Fixed only when that check no longer finds the problem. A problem that one page cannot decide says so and waits for the next full audit. Each site has an hourly allowance of page checks.
Read or export a report
Open Site report for uptime, certificate expiry, Lighthouse lab scores, search totals and audit findings. Choose Short summary for the site owner for the first page a site owner reads: what to do next, the main numbers and the five most serious findings. Full report adds the daily uptime, timings, search tables and every finding. Expand the history sections for response times, Lighthouse runs and recent outages. Use Print or save PDF for the report, or Download CSV for findings and search rows. Unknown means Hosting has no measurement for that item.
Connect a search provider
Open Search connections on the site's card.
- Google Search Console: open Settings, then Users and permissions, and choose Add user. Add the
displayed service account address with Restricted access. Enter the property as Search
Console shows it, for example
sc-domain:example.comorhttps://www.example.com/. - Bing Webmaster Tools: open Settings, then API access, and choose API Key. Enter the verified site URL with its trailing slash and the key. Enter the key again if you reconnect.
Check Imported through to see how recent the numbers are. Google data often trails by about three days, and Bing reports weekly. A new connection shows Pending first import until data arrives. Custd retries temporary import failures. If access was revoked, restore it at the provider and select Reconnect. Confirming Disconnect also removes the imported numbers for that provider. You can disconnect an existing provider while the domain awaits verification.
Share a live report
Open Share report on the site's card. Choose an expiry between 1 and 90 days, then select Create public link and copy the link before closing the section. You can see its creation and expiry dates later. Anyone with the link can read the current report. Use Revoke link to stop access before it expires, even if the report cannot load or the domain awaits verification.
Where your code and files live
An environment holds two different kinds of file, and they live in different places.
- Your application's code comes from your repository. Each release is built into an immutable bundle and
mounted at
/appin the application's container. Every release gets its own bundle, and earlier ones are kept, so a rollback runs exactly what ran before. You change code by committing and deploying — never by editing files inside the running container, which the next release replaces. - Your application's files — uploads, user content, anything written at runtime — live on the environment's files volume, shared by its processes.
The Files tool browses the second one, not the first. If your application is built from a repository and writes no user files, an empty Files listing is the correct answer: nothing has been written there. Your code is not missing — it is inside the release, which the file browser does not show.
Encryption at rest
Hosting supports encrypted file and database volumes: an environment's files volume and database storage are provisioned from the platform's encrypted storage class, so the bytes written to the platform's disks are encrypted rather than plaintext. The protection is at the volume layer, which has consequences worth knowing:
- It covers the stored volume, not every copy of your data. Backups rest on their own archive format and repository password (backups), and anything you export or copy keeps the protection of wherever it lands.
- It is not a backup. A restore point is still the only answer to data that was deleted or changed.
- We hold the key material for these volumes. You do not manage a separate key, so a copy you take out of the platform is not encrypted by us.
- Object storage you buy separately has its own protection; see object storage.
Provisioned storage is an allocation, not consumption
The Storage figure under Settings → Resources is space set aside for volumes, not what is written on them. A volume with nothing on it reads the same as one that is nearly full. For what each volume actually holds, open Settings → Storage: Storage on each volume shows what is Provisioned, Used and Free on each.
Extra storage
Object storage you add is bought separately from that allocation, in 10 GB blocks per month. Each month includes
downloads equal to the capacity you bought, and anything above it is charged per additional 10 GB.
Purchase it yourself from the company's storage page, or with hostingctl storage create. The panel and the
CLI both show the price from the price book before you confirm, and usage the platform cannot measure is
reported as unknown rather than as zero.
What sits in front of your site
Requests reach your application through the platform's edge, not directly:
- plain HTTP is redirected to HTTPS;
- responses are compressed with brotli, gzip or zstd, whichever the visitor's browser accepts, so text, JavaScript and CSS travel considerably smaller. You do not configure this, and nothing is compressed twice — a response your application compressed itself is passed through untouched.
Tools
The Overview's Tools view opens the environment's browser tools:
| Tool | What it is for |
|---|---|
| Files | Browse and edit the environment's application files |
| Database | Run SQL against an application's database (Adminer) |
| Terminal | A shell inside the environment |
All three open already signed in with your Hosting account, in their own browser tab. Choosing the same tool again focuses the tab you already have instead of opening a second one, so you do not lose work in it. If your browser blocks the new tab, allow pop-ups for Hosting and try again.
The tools reach the whole environment. A browser tool that can run SQL or edit files is as powerful as the credentials it uses, so treat access to it as administrator access.
Troubleshooting
Troubleshooting shows the technical detail behind the state: the environment's identifier, the generation that has been applied, each workload's instance count and allocation, storage claims and databases. It is there for when you are working something out with us, and it changes nothing.
Settings
Settings holds what belongs to the whole environment rather than one application. It has four views: Resources, Budget, Storage and Cancellation. The short version of changing resources is Change resources.
Resources: one shared pool
Your applications share one pool of what you bought. Each application can use whatever the pool has free at that moment; nothing is set aside for it unless you reserve it.
Your pool shows, for CPU, Memory and Storage:
- Bought — what the environment's purchase includes;
- Set aside for your databases — what your database servers take first, before the applications;
- Reserved by applications and services — the total of every application's reservation (for storage, Set aside for your volumes);
- Left — what is still free to reserve. A negative figure is shown in red and refused when you apply.
What reserving means. A reservation is a guaranteed share, not a cap. When the pool is quiet, every application can use far more than it reserves. When every application is busy at once, each still gets what it reserved. Reserve more for an application that must stay responsive under load; leave the rest unreserved so whichever application is busy can use it. The pool never reserves more than is left after the databases.
There is no per-application limit in the pool, because a limit stops an application using spare capacity it needs; Limit this application (not recommended) explains this beside each application. Each application runs as one instance in the pool. Your WordPress sites share one WordPress runtime and are sized together.
Database servers sit beside the pool. Each one is sized per instance: by default Hosting sets its CPU and memory (the row says what Auto sets), and Set this allocation myself lets you choose them and its instance count.
Changing it. Choose Change resources, adjust what each application reserves, then choose Apply changes. Review your changes lists every change, and Confirm and apply applies it. Every application in the pool restarts once while the change is applied, so sites may be briefly unavailable. Each application's own Resources tab shows what it reserves and, under Measured usage, what it has actually used.
Budget: reserved beside used
Budget lists each part of the environment that reserves from what you bought: the shared applications, each application that runs Pods of its own (one with two or more instances, or your own image), and each database server you pay for. For each it shows:
- Reserved — the CPU and memory it holds, across all of its Pods;
- Used — its CPU at the 95th percentile and its memory peak over the last 14 days;
- Suggestion — Could be smaller, Fits or Needs more, with the size and the reason.
A suggestion is only made from two full weeks of measurements. A part with less shows Not enough
data yet and says how much was measured, so a quiet or busy stretch never decides it. Memory is
sized from its peak, because a process past its memory limit is stopped; a short CPU burst is
served by the CPU limit and does not raise the suggestion. Nothing changes until you resize under
Resources. The same view is hostingctl environment budget <uuid> --company <slug>.
Storage and Cancellation
Storage shows what each volume holds and the Storage type for new volumes. Cancellation is the way out of the environment, with its current status — see Cancelling hosting.
Create an environment
The steps are in Create an environment: give it a name, choose a permanent slug, and pick one of the starting packages the platform offers. Placement is HaakCo's: you never choose a cluster. The package is the whole starting point — the purchase, the applications, the database and the tools — so you never assemble resource numbers yourself. The slug is reserved permanently because it addresses the environment's secret folder, so reusing it would hand a new environment the previous one's credentials.
Choose Empty Environment to reserve 2 CPUs, 4 GiB memory and 25 GiB storage without
creating applications, database servers, caches or volumes. Add only the services your application
needs afterwards. Its API package key is empty_environment.
Ten small sites via the API
The same creation is available through the company-authorized API. Read the packages the server offers, then create from one by its key:
GET /api/v1/c/{company}/hosting/environment-packages
POST /api/v1/c/{company}/hosting/environments
{
"name": "Small sites",
"slug": "small-sites",
"resourceProfile": "small_sites"
}
The server composes the budget, workloads, volumes and databases from the package, so name and the
permanent slug are the only fields the request adds. A budget that disagrees with the platform's
purchase is refused rather than applied.
Placement is chosen by HaakCo. The request carries no cluster, and sending a hostingClusterUuid
is refused as an error. An installation that has not named its placement target refuses creation
rather than guessing.
small_sites starts with the standard package's resource allocation, one database server and one
Redis. Use Add application to install each WordPress site with its own files, database and
administrator. The standard hosting sizes have no fixed site-count
limit: CPU, memory, storage and each application's requirements determine what fits. Only the Low End
package has a fixed application-count limit.
The database, tools and backup reservations come out of that one purchase, and the sites share one runtime and its costs inside your environment, never with another company. Resources are shared, not private allowances per site: a busy site can slow its siblings, and restarting the shared runtime affects them together. Larger or consistently busy sites need more resources or separate environments. The current quote shows purchased resources and price; a package is not a performance guarantee.
Creation records the desired state and applies it in the background. Check Applied and each application's serving state before treating it as ready. Use Settings and the application and address controls afterwards.
Next
- Open a site or an administrator login: Applications.
- Change what each application reserves: Change resources.