---
metadata:
  - name: generator
    content: Diplodoc Platform v5.62.0
alternate:
  - https://ydb.tech/docs/en/devops/observability/diagnostics.md?version=main
  - https://ydb.tech/docs/ru/devops/observability/diagnostics.md?version=main
  - href: https://ydb.tech/docs/en/devops/observability/diagnostics.md?version=main
    type: text/markdown
    title: Markdown version
  - href: https://ydb.tech/docs/en/llms.txt?revision=5f9509b7856960211eefd47b7217d87c71bdf904&version=main
    rel: describedby
sourcePath: en/core/devops/observability/diagnostics.md
---
> **Documentation Index:** Fetch the complete configuration index at https://ydb.tech/docs/en/llms.txt?revision=5f9509b7856960211eefd47b7217d87c71bdf904&version=main

# Diagnosing an isolated cluster

If a cluster is hidden behind network barriers, direct monitoring of its state is impossible. In such cases, you can use a mechanism for saving its state to a file to detect issues in the cluster's internal environment.

## Collecting diagnostic information

To collect diagnostic information, use the command:


```bash
ydb [global options...] admin cluster diagnostics collect \
  --duration 200 \
  --period 15 \
  --output diagnostics.tar
```


### Command parameters

* `--duration` — the number of seconds during which information about the cluster will be collected.
* `--period` — the interval in seconds between metric collections.
* `--output` — the path and file name in `.tar` format, to which all collected cluster information will be written.
* `--no-sanitize` — enables a full report by disabling the removal of user table, column, and query names (which are stripped by default).

The command writes the internal state of the cluster nodes to the specified file over a specified period. Every `period` seconds, metrics are collected from each cluster node and written to the file.

## Archive contents

The resulting file is an archive and contains:

* Metrics collected from the cluster nodes.
* Cluster information.

All data is written in JSON format in a human-readable form.

### Cluster information

Cluster information includes:

* Health Check report.
* Queries to the tables of the cluster partition `/Root/.sys` and their responses.
* Current configuration of the cluster subsystems.
* Current state of the cluster nodes.