====== Edit a collector ======

The **Edit Collector** modal opens from the **Registered collectors** table (Edit row action)

It groups every collector setting in one place, with a left-rail navigation by topic

You can switch sections back and forth before saving - unsaved changes are kept until you click **Save** or **Cancel**

===== Modal layout =====

^ Element ^ Description ^
| **Header** | Shows the collector name being edited |
| **Left rail** | Section navigation |
| **Main area** | The form for the current section |
| **Footer** | Cancel + Save buttons |

{{..:agents:edit_collector.png?800|}}

===== Section: General =====

Identity of the collector Saved through the standard agent update channel

^ Field ^ Description ^
| **Active** | Enable or disable the collector |
| **Short name** | Read-only abbreviated identifier (set at registration time) |
| **Title** | Full display name Required |
| **Description** | Optional free-text |
| **Properties** | Custom key/value table Useful for tagging collectors per environment, team, or region |

===== Section: Logs =====

Controls the Collector's log output 
^ Field ^ Description ^
| **Log level** | DEBUG / INFO / ERROR |
| **Debug log level expiration** | How long the DEBUG level stays active before falling back to INFO Only enabled when Log level is DEBUG Prevents leaving the Collector in DEBUG forever |

===== Section: Alerts =====

Alarm thresholds the Collector uses to raise alerts on its own health Two subsections

==== System availability alerts ====

Thresholds about connector reachability and timing

^ Field ^ Description ^
| **Max connection resp time (sec)** + severity | Connector is slow to answer beyond this time → raise alarm at the chosen severity |
| **Max system down time (sec)** + severity | Connector has been down for longer than this → raise alarm |
| **Unknown state timeout (sec)** + severity | Connector state stays unknown for too long → raise alarm |
| **Time zone alarm** severity | Severity used when the connector reports a wrong timezone |

==== Internal alerts ====

Thresholds about the Collector's own monitor execution

^ Field ^ Description ^
| **Monitor job execution error** severity | Severity raised when a monitor job throws |
| **CCMS errors** severity | Severity raised on CCMS-level errors |
| **Monitor Tree loading errors** severity | Severity raised when the monitor tree fails to load |

===== Section: Workers =====

Collector engine sizing Change with care - these settings affect Collector performance

^ Field ^ Description ^
| **Threads** | Parallel monitor execution threads |
| **Monitoring processes' memory (GB)** | JVM heap allocated to the worker pool |
| **Monitoring processes' timeout (min)** | Kill a monitor job that takes longer than this |
| **Do not compress results** | Toggle If ON, monitor results are not gzipped before they hit the storage backend Saves CPU but uses more disk and network |

**Important:** Increasing threads or memory beyond the host's capacity can crash the Collector Match the host's CPU and RAM

===== Section: Retention =====

How long the Collector keeps its various data stores before pruning

==== Storage toggles ====

^ Field ^ Description ^
| **Activate monitors results retention** | Turn the monitors-results retention job on/off |
| **Store statistics only** | If ON, the Collector stores only the per-monitor statistics, not the full results |
| **Store aborted monitor results only** | If ON, only failed/aborted job results are kept - successful runs are dropped |
| **Send monitor statistics metrics** | If ON, the Collector forwards per-monitor stats to the pipeline (alarms, plugins) |

==== Retention durations ====

^ Field ^ Description ^
| **Store jobs with duration above (sec)** | Only keep monitor results whose execution time exceeds this value - useful to drop fast-passing runs |
| **Monitors results retention (days)** | How long monitor results are kept |
| **Tests retention (days)** | How long manual test results are kept |
| **Log files retention (days)** | How long Collector log files are kept |
| **Reports retention (days)** | How long generated reports are kept |
| **Reports retention (max cleanup)** | Hard cap on the daily cleanup batch size for reports |
| **Snapshots retention (max)** | Max number of snapshots kept |
| **Configuration changes retention (days)** | How long the audit trail of configuration changes is kept |

After the retention window, the Collector deletes data on a daily cleanup job

===== Section: Licenses =====

^ Element ^ Description ^
| Info banner | //This Collector manages its licenses locally - use the Collector's own UI// |
| **Max expiration delay (days)** | Threshold for the //license about to expire// alarm Days before expiration at which the Collector raises a warning |


===== Section: Telnet =====

Settings for the Telnet check used during troubleshooting

^ Field ^ Description ^
| **Activate telnet check** | Master switch (toggle) |
| **Telnet timeout (sec)** | Idle session timeout |

===== Section: Reports =====

Scheduled report timeout

^ Field ^ Description ^
| **Max timeout report (sec)** | Maximum time a single report generation is allowed to run before it gets killed |

===== Section: OS Agent =====
Configuration of OS metrics ingestion from Telegraf agents 

This section is the main entry point for the OS Agent feature It contains four tabs:

^ Tab ^ Purpose ^
| **API Keys** | Global key and per-system overrides |
| **Configuration** | Input plugins, processes, generated setup files |
| **Hosts** | List of pushing hosts with status and detail |
| **Statistics** | Push counters |

{{:products:cockpit:1.0:userguide:settings:agents:agentos.png?600|}}

For the full OS Agent documentation, see [[products:promonitor:latest:userguide:administration:os-agent|OS Agent]] and its sub-pages