# System logs

> Find the warnings and errors behind a problem, from the hub and every edge gateway, and set how much Proxus logs and for how long.

Source: https://proxus.io/docs/administer/system-logs

At the end of this guide, you have found the log entries that explain a problem, such as a device that stopped
sending data or a rule that did not fire. The hub and every edge gateway send their warnings and errors to one list,
so you do not have to sign in to each machine.

## Before you start

- A user whose role may open **Platform > System Logs**.
- An idea of when the problem started and where (a gateway, a device, a rule).

## Find the entries

Go to **Platform > System Logs**.

*Screenshot: The System Logs list searched for Gateway, with the view selector, the Source and Count columns and the Support Bundle button marked*

1. The view selector: **Problems** (warnings, errors and critical entries, the default), **Errors** (errors and
   critical entries) or **All logs** (every level).
2. **Source** says where the entry comes from: **Platform** for the hub, **Gateway 1** and so on for each edge
   gateway (by its ID).
3. **Count** is how often the same entry happened. Repeats are shown once, with the time of the latest one.
4. **Support Bundle** packs the logs for Proxus support. See [Support bundle](https://proxus.io/docs/administer/support-bundle).

To narrow the list:

- Type a device name, tag name, rule name or gateway in **Text to search...** and press Enter.
- Use the filter icon in a column header, for example to show one **Level** or one **Source**.

**Reason** is a short code for the kind of problem, where Proxus knows one, for example a notification limit that
was reached.

## Read an entry

The list cuts each message to one line. Select an entry to open it.

*Screenshot: A log entry opened from the list: the level, time, source, reason and count on top, the whole message with its Copy button, and the identifiers below*

1. The level, the time of the latest occurrence, the source, the reason and how often it happened.
2. The whole message. **Copy** puts it on the clipboard, for example for a support case. When the entry has a
   **Diagnostic** (more technical detail, see `IncludeDiagnosticDetails` below), it follows with its own **Copy**.
3. **Identifiers**: the category, event ID, correlation ID, fingerprint and gateway ID of the entry. Proxus support
   may ask for them.

**Close** returns to the list.

The list shows the last day by default and at most 1,000 groups of entries; the footer says how many groups are
shown. To look further back, change `LogsTimeRange` in the `[Data]` section of the
[configuration file](https://proxus.io/docs/administer/configuration), for example `LogsTimeRange = "7d"`.

## How much is logged

`Loglevel` in the `[Logging]` section sets the lowest level that is stored:

| Value | Level |
| --- | --- |
| `0` | Trace |
| `1` | Debug |
| `2` | Information |
| `3` | Warning (default) |
| `4` | Error |
| `5` | Critical |

With the default, only warnings and worse are stored, so **All logs** shows the same as **Problems** most of the
time. Lower the level only while you look for a problem, then set it back: lower levels write much more.

> **Note:**
The shipped `docker-compose.yml` sets `Logging__Loglevel: "3"` on the hub container. This environment variable
wins over the file. To change the hub's level on Docker, change it in the compose file and recreate the container.

Other keys in `[Logging]`:

| Key | Default | Meaning |
| --- | --- | --- |
| `OperationalLogDedupeWindowSeconds` | `60` | The same warning or error within this many seconds is stored once, then every tenth time. Each stored entry carries how many times it happened since the one before, so **Count** in the list is the real number. |
| `IncludeDiagnosticDetails` | `false` | When `true`, entries keep more technical detail. Turn it on only for a support case. |
| `LogToFile` | `true` | Also write log files on disk (see below). |
| `LogSqlQueries` | `false` | Log every database query. Only for support cases: it writes a lot and can contain data. |

## How long logs are kept

| Where | Kept for | Set by |
| --- | --- | --- |
| The **System Logs** list (telemetry database) | 45 days | `ClickHouseLogTTLExpression` in `[ClickHouse]` |
| Log files on disk | Until you delete them | One file per day, a new one after each 1 MB |

The log files are in the configuration folder (`/app/config` in the containers), named by date, for example
`2026-10-15_000.log`. They are part of the `config` volume. Delete old ones from time to time, or keep them small by
leaving `Loglevel` at `3`.

## What's next

- [Send the logs to Proxus support](https://proxus.io/docs/administer/support-bundle)
- [Troubleshoot](https://proxus.io/docs/troubleshoot)
- [Monitor the installation](https://proxus.io/docs/operate/monitoring)