Database Explorer
Every database on every server DBackup reaches, with the jobs that back it up, and its tables and rows read live without a separate database client. The servers themselves come with every version they ran.
Overview
The Databases tab lists every database of every database connection in one table:
- The server, the size and the number of tables of each database
- Backed up by names the enabled jobs that back it up, and In no job marks a database nothing backs up
- Last backup says when a kept backup last held it
A Redis or Valkey server is one entry with the keys of all its numbered databases, since DBackup backs it up as a whole. Its line says how many of its databases hold keys.
The numbers above the list count the databases, their size, their tables, how many are in a job and how much data no job backs up. The keys of Redis and Valkey do not count as tables.
The filters for the server, the engine, the job and the state count what the other filters leave. A value without a database under them moves to the end of its list and cannot be picked, so with PostgreSQL picked as the engine only the PostgreSQL servers are left to pick.
Where the Data Comes From
The list does not ask a server when the page opens. DBackup reads the databases of every server with their sizes once an hour, in the same run that checks the version (the Update Database Versions system task). The button in the top right says how old the lists are, and its Read now reads every server at once.
A server whose read fails keeps the list from before, and the button names it with the error. Everything else on the page comes from DBackup itself: the jobs, the backups of each database from the runs, and the versions from the version history.
Tables and rows are the only data read live, when you open them.
Table and Timeline
The switch in the top right shows the list as a table or by day, like the Backups page. A phone shows the databases as cards and has no switch.
The timeline groups the databases by server, with a column per day:
| Mark | Meaning |
|---|---|
| Gray cell | A backup held the database that day, with the number of runs when there were several |
| Red cell | A run that should have backed it up failed |
| Gray cell with a red dot | A run failed and another one backed it up that day |
| Dashed cell | A run the schedule plans, shown for the next 7 days with the arrow at today |
| Arrow with a version | The server runs a new version since that day, the dashed line marks the day over its databases |
A server updated on several days in a row gets a mark on each day, and several updates on one day show the newest with a count. The hover of a mark lists each change with the time it was read.
WARNING
A backup is only restored onto the same version or a newer one. The backups made after an update need a server with that version.
The server heads name the server on top and its engine, version and how many of its databases are in a job below. A server with more than 10 databases starts folded into one row with every run of it, and its arrow opens it. The rows come in pages like a table, with the rows per page of your profile.
The Backups of a Day
A click on a day shows its backups in a panel beside the timeline, or over it on a smaller screen. The dropdown in its top right lists the runs of that day by time, however many there are, and the panel shows the one picked:
- A backup shows like on the Backups page, with its copies, what it holds and its integrity. Restore starts with only the database you clicked ticked, the menu restores the whole backup, and Download, Lock and Verify work as there
- A failed run shows its error with a link to History, and the backups before and after it, which Show opens in the panel
- A run the schedule still plans that day offers Run now
- A backup no destination holds anymore, like after retention removed it, says so and points to the backups around it
A click on a name opens the page of the database with its live tables.
The Page of a Database
A click on a database opens it as a page of its own. The arrow in the top left leads back to the list, and the field beside the name opens another database of any server.
- The numbers: tables, size, the jobs that back it up, the last backup and the version of the server with the one before
- Open backups shows its backups on the Backups page, Open job the job that backs it up
- The tables, biggest first, with a search, and the rows of the picked one beside them
On a wide screen the tables and the rows fill the window and each scrolls on its own. The picked table is part of the address, so a link opens it again.
A click on a table reads its rows from the server, 50 at a time:
- Filter looks through the whole table for a value in one column, as contains, is, starts with or ends with
- A click on a column head sorts by it, a second one reverses the order and a third removes it
- The columns button hides columns, and the Columns tab lists how the table is built
- MongoDB shows documents
Redis and Valkey
The page of a Redis or Valkey server lists its numbered databases with their keys. The empty ones are folded into one line at the end, which a click opens.
A click on a database reads its first keys with the type and the time to live. Filter matches the key on the server, as contains, is, starts with or ends with. It looks through up to 50,000 keys and shows the first 200 that match. The refresh button beside Databases counts the keys again.
The explorer only reads. It runs no queries of its own and changes nothing.
The Servers Tab
The Servers tab lists every database server with:
- Where it runs, as the host and port of its connection
- Its version, since when it runs it and the one before
- How many of its databases a job backs up, as a bar with the ones in no job in amber
- The size of its databases, and its kept backups with the time of the last one
- Its status with the response time of its last health check
A phone shows the servers as cards. The numbers above the list count the servers, their databases, the new versions of the last 30 days, the servers behind and the ones online. The quick filters keep the servers Behind or Not all backed up, and the Engine filter keeps one kind of database. It counts the servers the quick filter leaves.
A server is behind when a kept backup of another server with the same engine has a newer version than the one it runs. A backup restores only onto the same version or a newer one, so those backups would not restore onto it. The line under its version names the server and the version it is behind.
The Page of a Server
A click on a server opens it as a page of its own. The arrow in the top left leads back to the list, and the field beside the name opens another server.
- The numbers: its version, its databases or keys, their size, its kept backups and how often it answered the health check in the last 30 days
- Open backups shows the backups of its jobs on the Backups page, Open connection leads to the Connections page
- A server that is behind says so on top, with the version its backups would need
Versions lists every version the server ran, newest first, five a page:
| Column | Meaning |
|---|---|
| Version | The version with Now on the current one, and the one before it with an arrow. A step back to an older version is marked in amber |
| Since | When the hourly version check found it |
| Until | When the check found the next one, or today |
| On it | How long the server ran it |
| Backups | How many backups its jobs made while the server ran it, and how many of those a destination still keeps |
The oldest version is the one the server ran when it was added to DBackup. Retention removes old backups, so an older version often shows backups made and none kept.
Databases lists the databases of the server, biggest first, with the jobs that back up each. A click opens the page of a database, and All in Databases shows every one of them on the Databases tab. A Redis or Valkey server lists its numbered databases with their keys.
Required Permissions
| Permission | What it grants |
|---|---|
sources:view | The page, the databases and their sizes, the servers and their versions |
sources:read | The tables and rows of a database |
jobs:read | The jobs of each database, its last backup and the timeline, and the backups made on each version of a server |
storage:read, storage:restore, storage:download | The backup of a day in the panel, restoring it and downloading it |
jobs:execute | Run now for a day the schedule still plans |
storage:read | Open backups, the kept backups of each server and which servers are behind |
history:read | The link from a run to History |
A Login That May Only Back Up
The explorer reads with the login of the connection. For MySQL, MariaDB, PostgreSQL and MongoDB a backup login may read every row anyway, since a dump reads every table, so whoever has sources:read sees the data. That permission is what keeps rows private.
SQL Server with db_backupoperator and a Redis user limited to +sync and +psync can back up without being allowed to look inside. The explorer then lists the database with its jobs and backups, shows a dash where the size is not told, and says in place of the tables or rows why they stay closed. The backups are not affected.
Troubleshooting
A server is missing its databases
Open the button in the top right. It names every server whose last read failed with the error of the server. Fix the connection, then use Read now.
A database shows no tables
The login of the connection may not see them. Grant it the rights to read the database, or leave it: the jobs and backups of the database still show.
The timeline is missing
The timeline needs the jobs:read permission. Without it the list shows the databases without their jobs.
Next Steps
- Database Sources - Configure and manage database connections
- Backups - Every backup across destinations
- Groups & Permissions - Assign
sources:readandjobs:readto groups