Compression

Reduce backup file sizes with Gzip or Brotli compression.

Overview

Compression significantly reduces:

• Storage costs - Smaller files = less storage
• Transfer time - Less data to upload
• Network bandwidth - Lower bandwidth usage

Typical compression ratios for SQL dumps:

Content Type	Compression Ratio
SQL dumps (text)	60-80% smaller
Binary data (BLOBs)	10-30% smaller
Already compressed	0-5% smaller

Compression Algorithms

Gzip

Best for: General use, fast compression

Aspect	Rating
Speed	⭐⭐⭐⭐⭐ Fast
Compression	⭐⭐⭐ Good
CPU Usage	Low
Compatibility	Universal

Brotli

Best for: Maximum compression, slower systems acceptable

Aspect	Rating
Speed	⭐⭐⭐ Moderate
Compression	⭐⭐⭐⭐⭐ Excellent
CPU Usage	Medium
Compatibility	Modern

DBackup runs Brotli at quality 10 rather than its maximum of 11. On a measured 22 MB SQL dump that costs 2.5% in size and returns well over half the time. Quality 9 and below are a different setting rather than a faster version of the same one - they skip the search that makes Brotli worth choosing, and produce output around 85% larger. If that trade appeals, Gzip already offers it.

Comparison

Indicative figures for a 100 MB SQL dump. Your ratio depends entirely on the content, and Brotli's lead over Gzip grows the more repetition there is:

Algorithm	Compressed Size	Relative Time
None	100 MB	-
Gzip	25 MB (75% reduction)	1x
Brotli	20 MB (80% reduction)	15-25x

Enabling Compression

On Job Creation

1. Create or edit a backup job
2. Open its Compression part
3. Pick Gzip or Brotli. A PostgreSQL job lets pg_dump compress the dump under The dump, and DBackup only compresses that dump when it is set to None.
4. Save

File Extensions

Every backup is a seekable archive that compresses each database dump and each file on its own, so the archive keeps its .tar name. In an unencrypted archive the entries carry the extension instead, for example databases/shop.sql.gz.

Backups written by earlier versions for database-only jobs are compressed as a whole and have extensions:

• Gzip: backup.sql.gz
• Brotli: backup.sql.br

With encryption:

• backup.sql.gz.enc
• backup.sql.br.enc

Already-Compressed Formats

Every file of a directory source is compressed on its own. Files whose format is already compressed are stored as-is instead, even when the job has compression enabled. The same goes for database dumps the engine already compressed: PostgreSQL custom-format dumps with compression on, MongoDB archives and Azure SQL BACPACs.

Recompressing them gains a fraction of a percent at best, and costs the full CPU time plus a complete extra write and read of the file through a temporary file. On a photo or video library that is the difference between a backup that finishes and one that does not.

The files are still in the backup

This only skips the compression step. Nothing is left out. To leave files out of a backup, use exclude patterns on the directory source.

Formats stored as-is

Matched on the file extension, case-insensitive, and only the last one - so photos.tar.gz counts as gz.

Category	Extensions
Video	3gp avi flv m2ts m4v mkv mov mp4 mpeg mpg mts ts vob webm wmv
Audio	aac ape flac m4a mka mp3 oga ogg opus wma
Images	avif gif heic heif jp2 jpeg jpg jxl png webp
Archives	7z br bz2 cab gz lz4 lzma rar tbz2 tgz txz xz zip zst
ZIP containers	apk bacpac docx epub ipa jar nupkg odp ods odt pptx vsix war whl xlsx xpi
Web fonts	woff woff2
Encrypted	age enc gpg pgp
Disk images	dmg

Everything else is compressed normally. Some formats look like they belong on this list and are left off on purpose, because they compress well often enough to be worth the attempt - PDFs, executables and ISO images among them.

If a format is missing and you are backing up a lot of it, that is worth reporting - the list ships in the code and grows with releases, without anything to reconfigure.

Behaviour

It is automatic and needs no configuration. The job log names how many files were affected:

412 file(s) stored uncompressed (38.1 GB) - their format is already compressed

Restore, download and file-level restore handle this by themselves, and so does the Recovery Kit, including kits you downloaded earlier. Every file records how it was stored, so nothing has to be worked out at restore time.

Unpacking an unencrypted backup by hand is the one place it shows: a stored file keeps its real name and is immediately usable, while a compressed one ends in .gz or .br.

Pipeline Order

Each database is dumped to a temporary file first. The archive writer then takes one entry at a time, compresses it into a temporary file of its own, and encrypts it while it writes it into the .tar. The finished .tar is uploaded last:

Database → Dump (temp file) → Compress (temp file) → Encrypt → Into the .tar → Upload
                                   ↑                    ↑
                              Gzip/Brotli          AES-256-GCM

Compression comes before encryption because encrypted data does not compress. Each entry is compressed on its own, never the .tar as a whole, so one database can be read out of the archive without the others. Every step streams, so memory use does not grow with the size of the database.

When to Use

Use Gzip When

• Backup speed is critical
• CPU resources are limited
• Compatibility is important
• Good balance needed

Use Brotli When

• Storage costs are high
• Maximum compression wanted
• Time is not critical
• Modern systems only

Skip Compression When

• Database contains mostly binary BLOBs
• Network is faster than compression time
• Immediate backups needed

Storage Savings

Example: 1GB Database

Setup	Size	Monthly Cost*
No compression	30 GB (30 daily)	$0.69
Gzip	7.5 GB	$0.17
Brotli	6 GB	$0.14

*S3 Standard pricing ($0.023/GB)

Yearly Savings

For 10 databases with Smart retention:

• No compression: ~300 GB = $83/year
• With Gzip: ~75 GB = $21/year
• Savings: $62/year per 10 databases

Restore and Download

Automatic Decompression

When restoring or downloading:

1. DBackup reads metadata
2. Detects compression algorithm
3. Decompresses automatically

Manual Decompression

If needed outside DBackup:

# Gzip
gunzip backup.sql.gz

# Brotli
brotli -d backup.sql.br

Performance Tuning

CPU Considerations

Compression uses CPU. Monitor during backups:

• High CPU: Consider Gzip over Brotli
• Multiple jobs: Stagger schedules

Memory Usage

Streaming keeps memory low, but:

• Brotli uses more memory than Gzip
• Large databases may need more resources

Disk I/O

Compression adds disk work, since each entry is compressed into a temporary file before it goes into the archive. In return the archive and its upload are smaller.

Metadata Storage

The .meta.json beside the archive names the algorithm under archive.compression, left out when the job compresses nothing. What restore reads is the comp field of each entry in the archive's index, absent for an entry stored as is. Backups written by earlier versions for database-only jobs, compressed as a whole, keep a top-level compression field instead. See Archive Format for the layout.

Troubleshooting

Compression Slow

Causes:

• Large database
• Brotli algorithm
• CPU constraints

Solutions:

1. Switch to Gzip
2. Schedule during low-usage
3. Check CPU availability

Decompression Fails

Causes:

• Corrupted file
• Wrong algorithm detected
• Incomplete download

Solutions:

1. Re-download from storage
2. Check .meta.json for correct algorithm
3. Verify file integrity

File Larger After Compression

Cause: Already compressed data. In a file backup, known formats are stored as-is automatically, so this points at content the extension does not give away - a database dump full of BLOBs, or files with a misleading name.

Solution: Set the job to None. There is nothing to gain from compressing data that is already compressed, and it costs the CPU time either way.

Best Practices

1. Start with Gzip - Good balance for most cases
2. Monitor backup times - Switch if too slow
3. Compare sizes - Test both algorithms
4. Enable for all jobs - Storage savings add up
5. Combine with encryption - Compress then encrypt
6. Test restores - Verify decompression works

Next Steps

• Encryption - Encrypt compressed backups
• Creating Jobs - Configure compression
• Backups - View backup sizes