A self-hosted Langfuse or SigNoz installation can fill its ClickHouse disk for three separate reasons: application data kept longer than you need, telemetry arriving faster than you planned for, or ClickHouse’s own system log tables. Each has a different control, and a retention setting only governs the data it covers. Lowering retention will not reclaim space taken by system logs. Measure first, identify the largest consumer, then apply the matching fix. Langfuse and SigNoz are handled separately below because no single setting addresses both.
The three sources in brief:
- Application data kept too long. In Langfuse, this is traces, observations, scores, and media. In SigNoz, it is logs, traces, and metrics, each governed by its own retention setting.
- High incoming volume. SigNoz’s production guidance says high ingestion volume increases ClickHouse storage. Retention does not reduce what arrives; only changes at the source do.
- ClickHouse system log tables. Langfuse’s FAQ says these can frequently dominate disk usage even after Langfuse retention is configured.
Measure before you change anything
Check disk capacity and growth rate
Start with the filesystem or persistent volume that holds ClickHouse data. On the host, run df -h against the mount point, or read the volume’s usage metric in your orchestrator. Then ask ClickHouse what it sees:
As an Amazon Associate I earn from qualifying purchases.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsSELECT
name,
path,
formatReadableSize(total_space - free_space) AS used,
formatReadableSize(free_space) AS free
FROM system.disksRecord used space at two points several hours apart and calculate the daily growth. The growth rate, more than the absolute size, tells you whether you have hours or weeks before the disk is full.
Rank tables with Langfuse’s active-parts query
Langfuse documents this query on its scaling page. It sums the bytes and rows of active parts for each table and database:
#1 Best Overall
- Easily store and access 2TB to content on the go with the Seagate Portable Drive, a USB external hard drive
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop
- To get set up, connect the portable hard drive to a computer for automatic recognition no software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
SELECT table, formatReadableSize(size) AS size, rows FROM (
SELECT
table,
database,
sum(bytes) AS size,
sum(rows) AS rows
FROM system.parts
WHERE active
GROUP BY table, database
ORDER BY size DESC
)Two adjustments may be needed depending on your ClickHouse version. If it rejects the reused size alias in the outer SELECT, rename the outer alias to size_readable; the logic is unchanged. The ORDER BY sits in the inner query, so if the output is not sorted, add ORDER BY size DESC to the outer query as well.
Read the result by table origin. If the largest rows come from the system database, the consumer is ClickHouse’s own logs. If they are application tables such as traces, observations, or scores, the Langfuse retention section below applies. The query counts only active parts, so it will not show data that has been merged away but not yet removed from disk. Do not assume the largest table is the same in every installation.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →SigNoz: locate the largest tables with the diagnostics for your version
The SigNoz guidance establishes its retention controls and ingestion-volume recommendations, but it does not provide a SigNoz-specific largest-table query. Inspect ClickHouse table and disk metrics using the diagnostics that match your installed ClickHouse version and deployment. Validate table names against your own schema before acting on them. Do not reuse Langfuse table names or Langfuse SQL on a SigNoz installation, because the two products use different schemas.
Langfuse: separate product data from ClickHouse system logs
Langfuse uses ClickHouse as its primary analytical store for traces, observations, and scores. Its disk growth has two sources: the data Langfuse writes, which is governed by Langfuse retention, and ClickHouse’s own logs, which Langfuse does not read and which need separate handling.
Rank #2
- Easily store and access 5TB of content on the go with the Seagate portable drive, a USB external hard Drive
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop
- To get set up, connect the portable hard drive to a computer for automatic recognition software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
Confirm the Langfuse and ClickHouse versions first
The Langfuse ClickHouse guide sets the following requirements:
| Langfuse version | Minimum ClickHouse | Recommended ClickHouse |
|---|---|---|
| Langfuse v3 | 24.3 | Not stated |
| Langfuse v4 | 25.12 | 26.4 |
Confirm both versions before recommending a version-dependent setting. Some deletion options described below require ClickHouse 25.7 or newer.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchSystem log tables that grow without limit
The Langfuse FAQ names six built-in ClickHouse system log tables: trace_log, text_log, opentelemetry_span_log, asynchronous_metric_log, metric_log, and latency_log. They ship without TTLs by default, and the query profiler writes to system.trace_log continuously. The FAQ states the central point directly:
On self-hosted Langfuse, disk usage is frequently dominated by ClickHouse’s built-in system log tables rather than Langfuse’s own tables.
Rank #3
Seagate Portable 1TB External Hard Drive HDD – USB 3.0 for PC, Mac, PlayStation, & Xbox, 1-Year Rescue Service (STGX1000400) , Black
- Easily store and access 1TB to content on the go with the Seagate Portable Drive, a USB external hard drive.Specific uses: Personal
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop. Reformatting may be required for Mac
- To get set up, connect the portable hard drive to a computer for automatic recognition no software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
The Langfuse scaling page describes two approaches:
- Disable the system log tables you do not use through ClickHouse server configuration.
- Keep them, but apply aggressive TTLs and disable the query profiler.
Keep query_log, part_log, and error_log enabled. The scaling page describes them as useful for debugging and as remaining small. Before you remove any system log table, confirm that your team does not use it for troubleshooting. The sources do not say which of these tables your installation needs.
If you disable a table through a configuration override, the usual pattern in ClickHouse is a file in the config.d directory that sets the table element with remove="1", for example <trace_log remove="1"/> inside <clickhouse>. Confirm the exact syntax and the handling of existing table data in your ClickHouse version’s documentation before applying it. To disable the query profiler, set the query_profiler_real_time_period_ns and query_profiler_cpu_time_period_ns settings to 0, again checked against your version’s settings reference.
Langfuse-owned data: project retention
Langfuse’s Data Retention feature deletes old traces, observations, scores, and media assets nightly, per project. Before you rely on it, check these prerequisites in the Langfuse retention guide and the scaling page:
Rank #4
- Easily store and access 4TB of content on the go with the Seagate Portable Drive, a USB external hard drive.Specific uses: Personal
- Designed to work with Windows or Mac computers, this external hard drive makes backup a snap just drag and drop
- To get set up, connect the portable hard drive to a computer for automatic recognition no software required
- This USB drive provides plug and play simplicity with the included 18 inch USB 3.0 cable
- The available storage capacity may vary.
- On self-hosted instances, the feature is available on Enterprise Edition.
- Without a configured policy, self-hosted event data is stored indefinitely. The feature is not a default cleanup job.
- The minimum retention window is three days.
- On self-hosted instances, the IAM role used for storage needs
s3:DeleteObjecton the buckets. - Versioned object storage may keep prior object versions or delete markers. These need separate cleanup.
- A saved dataset item can remain after its original trace expires.
- Do not apply a broad lifecycle rule to the media bucket to delete objects. Langfuse warns that this can break references and uploads.
ClickHouse-level TTLs and slow deletions
When application retention is not available to you, Langfuse documents ClickHouse TTLs for traces, observations, scores, and event_log. Use this only after you have settled the retention window you need and confirmed the product and plan limits above. It is a last resort, not a first diagnostic step.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Large deletions can run slowly. The scaling page lists LANGFUSE_CLICKHOUSE_DELETION_TIMEOUT_MS, with a default of 600000 ms (10 minutes). On ClickHouse 25.7 and newer, the page describes lightweight deletion options. Read that section before changing deletion behavior.
SigNoz: retention and incoming volume are separate controls
The documented default retention differs between pages
SigNoz’s first-party pages do not agree on default retention:
| Source | Logs and traces | Metrics |
|---|---|---|
| SigNoz Docker installation guide (current Docker and Kubernetes guides) | 7 days | 30 days |
| SigNoz retention guide | 15 days | 30 days |
Do not assume either figure is the default for your installation. Check the value shown in your deployed version’s settings.
Best Value
- [Upgraded Version] - This external hard drive features a mirrored logo stripe combined with a striped anti-slip design, and the rounded corners of the casing make it easier to grip. The stripes also have a heat dissipation function, ensuring stable and fast data transfer.
- 【Ultra-thin and quiet】 - The motherboard adopts JMicron 578 noise-free solution, giving you a quiet working environment. Lightweight and portable size designed to fit in your pocket for easy portability.
- 【Ultra-Fast Data Transfers】 - Pairing this external hard drive with JMicron 578 solution USB 3.0 and USB 2.0 interfaces enables blazing-fast data transfer. It boasts theoretical read speeds of up to 125MB/s and write speeds of up to 103MB/s.
- 【Plug and Play】 - With no software to install, just plug it in and the drive is ready to use.The hard disk chip is wrapped with an aluminum anti-interference layer to increase heat dissipation and protect data.
- 【What You Get】 - 1 x Portable Hard Drive, 1 x USB 3.0 Cable, 1 x User Manual, Gift-type shell packaging ,Three-year manufacturer's warranty and free technical support services.
Change retention for each signal
In the SigNoz UI, open Settings and look under Workspace. Some installation pages place retention under General, so use whichever section your version shows. Retention for logs, traces, and metrics can be set independently. Set each signal to the window you actually need, not to a single value copied from another installation.
Free tools Windows power users keep installed
One-click scans. No signup required.
Why disk space does not drop right away
A retention change applies only to newly ingested data. Existing data keeps the retention it was ingested with. TTL expiry deletes data permanently, and raising retention later does not restore data that has already expired. After you lower retention, space comes back gradually as older data reaches its earlier expiry, so a flat disk graph in the first days after the change is expected.
If you need older data removed sooner, the retention guide does not establish a supported procedure for applying a new window to existing data. Check the guide for your version before running manual deletes against ClickHouse tables.
Reduce incoming volume at the source
If growth is fast, the SigNoz production checklist recommends finding which signal and service drives the volume, then applying the matching control:
- Metrics: drop unused metrics and aggregate high-cardinality labels.
- Traces: sample traces.
- Logs: filter verbose logs in the OpenTelemetry Collector before they are stored.
- Long-term archival: the checklist documents an S3 exporter path. The archived copy is not queryable in the SigNoz UI, so it preserves data without replacing online retention. To reduce active storage, lower online retention separately.
Choose the fix by what it changes
Compare fixes on four axes: which data they affect, whether they reclaim existing data or only limit future growth, whether the data stays queryable in the product, and what they require.
Quick Recap
| Fix | Data affected | Existing data | Queryable in product | Requirements |
|---|---|---|---|---|
| Langfuse project retention | Traces, observations, scores, media | Deletes data past the window; physical space follows merges and object cleanup | No, for data past the window | Plan and object-storage prerequisites above |
| Langfuse system-log removal or TTL | ClickHouse system log tables | Removal: confirm handling for your version. TTL: rows age out | Not applicable; Langfuse does not read these tables | ClickHouse configuration access; version-specific syntax |
| SigNoz per-signal retention | Logs, traces, metrics | Not immediately; see the SigNoz section above | Until expiry | Checked against your installed version’s setting |
| SigNoz ingestion reduction | Future ingestion | No | Dropped or sampled data is not stored | Collector and pipeline configuration changes |
| SigNoz S3 archival | Long-term copy | No | No, not queryable in the UI | S3 exporter configuration and bucket access |
| Additional disk capacity | None; headroom only | No | Yes | Compatible storage interface, filesystem, and deployment topology |
Putting it together
- Check capacity and growth rate. Use the disk query above and the filesystem or volume metric. If the disk is near full, add capacity as immediate headroom. Adding capacity buys time and does not stop growth. A generic NVMe SSD can serve this purpose on a host that supports it, but confirm the slot, interface, and filesystem before buying hardware.
- Rank the tables. For Langfuse, run the active-parts query. For SigNoz, use the diagnostics for your installed ClickHouse version and validate the table names.
- Act on the Langfuse result. If application tables dominate, check that project retention is available to you and set the window you need. If system logs dominate, apply the system-log guidance and keep
query_log,part_log, anderror_logenabled. - Act on the SigNoz result. If telemetry tables grow quickly, check the retention for each signal, identify the service or signal that drives volume, and apply the matching ingestion control.
- Verify that deletion is taking effect. Repeat the disk query and the table ranking over time and compare the results with the growth rate from step 1.
- ClickHouse merges, TTL processing, object-storage deletion, and filesystem or volume capacity are separate measurements. Do not treat data as physically reclaimed until the volume’s free space rises.
- The active-parts query will not show merged-away data that is still pending removal from disk.
- Evaluate managed options if storage remains a bottleneck. The Langfuse scaling page names ClickHouse Cloud and bring-your-own-cloud (BYOC) deployments. The SigNoz production checklist documents Cloud-managed scaling and retention. This is an architecture and cost decision rather than a cleanup step, so compare cost and data-residency terms directly with each vendor.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
