Start free. Add your team later.

Always free for one user. The Team plan is $49 per user per month, and you can buy it now, also during the beta. To buy Team, or to ask about Scale, write to contact@guardioz.com.

One account. Today an organization has one account. Inviting more people comes to the Team plan in autumn 2026.

One PostgreSQL instance that you register with one install command.

Guardioz refuses a new server over your plan's limit when it registers, and the message says why. Servers you already registered keep being checked.

Yes. Create an account, copy the install command from the app, and run it on the server that runs PostgreSQL. No credit card is required, and the free plan needs no call with sales. The account is how Guardioz knows it is you: you confirm your email address, or sign up with Google or GitHub.

PostgreSQL 18 on a server that you run yourself, on Linux with systemd or on FreeBSD. A virtual machine counts. Guardioz installs on the primary server. Replicas are not supported, and neither are managed services such as Amazon RDS and Azure Database for PostgreSQL, PostgreSQL in a container, or clusters run by a Kubernetes operator. The full list is in the requirements.

The installer checks before it changes anything. On a PostgreSQL version other than 18, on a replica, or on Linux without systemd, it stops, says why, and leaves the install command unused.

PostgreSQL statistics, settings and system catalogs, and a profile of the machine: its memory, processors and disk. It never selects the rows of your tables. It runs a fixed set of read-only queries written in its source, in a read-only session, with a 30-second limit per database, and sends each snapshot out over HTTPS. Some of what it sends can carry your own text, such as the statements your sessions are running. The documentation lists every source and every such part.

Guardioz monitors your server around the clock, on a schedule. The agent runs as a service and collects a snapshot about once an hour without anyone starting it, and Guardioz checks each snapshot as soon as it arrives. The list of tables and indexes travels once a day, and their sizes on disk once a week. PostgreSQL's statistics are counters that keep adding up between two snapshots, so two snapshots an hour apart show the activity of the whole hour, not only one moment. The schedule is a design decision, made to keep the load on your database small while the checks still get the data they need. Guardioz is not real-time monitoring and does not draw second-by-second graphs of your metrics.

The agent is designed to keep it small. Every query a monitoring tool runs uses the same processors, memory and disk as your application. During an incident, when the server is already short of them, a tool that queries the database every few seconds adds to the pressure at the worst moment. The agent collects once an hour, and measures sizes on disk, its heaviest work, once a week. It runs its read-only queries one at a time, each with a time limit, stops waiting for a lock on a table after 5 seconds, and reads statistics and catalogs, never the rows of your tables. The documentation explains why the agent collects once an hour.

No. The agent starts the connection itself, outbound. You do not open an inbound port and you do not expose your database port to the internet.

Not on its own. The installer creates a read-only monitoring role and, only when that role cannot log in, adds one rule to pg_hba.conf. It changes no PostgreSQL setting. After that the agent only reads, and every change Guardioz recommends is a command you run yourself. For a per-table cleanup setting, you also get the command that restores the previous value.

There is no single uninstall command. Each step undoes one change that the installer made, and a file on your server records many of those changes, each with the command that reverses it. The documentation lists every step.

Right after you install the agent: the first snapshot gives a full result. Want to look before you install anything? The live demo needs no sign-up.

Monitoring shows how your metrics move and usually leaves the judgment to you. Guardioz reads your database's own signals and tells you which of them matter, how serious each one is, and what to fix first, with the measurements from your server that each verdict rests on.

A chatbot answers the question you ask, from what you paste into it. Guardioz keeps checking your server, snapshot after snapshot, including the problems you did not think to ask about, and shows the measured data behind each recommendation.

No. Guardioz is built for teams that run PostgreSQL without a DBA. Findings say what we found, why it matters and what to do next, and a settings change says whether it needs a restart.

No. A recommendation shows the measurements from your server that it rests on, so you can check it before you act, and you decide whether to run it.

Settings, indexes, table cleanup (autovacuum), replication slots, transaction log (WAL) archiving and open transactions. It does not analyze individual queries today. Query analysis is planned. The documentation lists every check.

Then you see that too: the checks that passed on your server are listed, not hidden behind an empty page. Guardioz keeps checking every new snapshot, and every check is in the free plan.

Guardioz is a product of Clouds on Mars. Read why we are building it.

Guardioz is in beta. You can sign up, and the product is early - we say so up front. We do not offer an SLA.

Write to contact@guardioz.com. The documentation says what to put in the message, so that we do not have to ask for more details.

Yes. Customers have a real say in what we build next - in the beta and after it. Write to contact@guardioz.com and tell us which problems matter most to you.