Server Health Indicator¶
The health indicator in the MMA main window summarizes the state of the local MetaMiner Server (MMS) from its logs, even when the server is down or will not start.
How it works¶
Every 5 seconds MMA reads, directly from disk on the server machine:
- the
metaminerWindows service state, - the incident files MMS writes to
server\health\(incidents-YYYY-MM-DD.jsonl), - the current
server\logs\server.log.
It does not need a connection to MMS, so it still works when MMS is stopped or failing to start. The indicator stays hidden until MMA has local data to show.
Image of the MMA main window with the health indicator showing an amber '3 notable' state
../images/admin/mma-health-indicator.png
States¶
| Colour | Text | Meaning |
|---|---|---|
| Red | server won't start | The last start attempt failed with a fatal error |
| Red | N urgent | N urgent issue types were seen recently |
| Amber | wedged (log not advancing) | The service runs but server.log stopped moving |
| Amber | N notable | N notable issue types were seen recently |
| Amber | stale | The health data is old |
| Grey | starting... / transitioning... | The service is starting or changing state |
| Grey | stopped | The service is stopped and no fatal error was recorded |
| Grey | no issues | Running, nothing urgent or notable |
Health details dialog¶
Click the indicator to open MetaMiner - Local Health.
| Element | Purpose |
|---|---|
| Table columns Urgency, Family, Count, Last seen, Host, Sample | One row per issue type (family) |
| Evidence pane | Full evidence for the selected row |
| Diagnose | Scans the tail of the local logs for the selected issue and shows a short analysis |
| Copy for Support | Copies the selected evidence to the clipboard for a support ticket |
| Refresh | Re-reads the health data |
| Close | Closes the dialog |
Image of the MetaMiner - Local Health dialog with one row selected and the evidence pane filled
../images/admin/mma-local-health-dialog.png
Diagnose runs only when you click it. The 5-second indicator refresh never scans whole log files.
Incident retention¶
MMS keeps incident files for 90 days and reloads the last 7 days when it starts.
Incident settings (system properties)
| Property | Default | Effect |
|---|---|---|
logshark.incidents.enabled |
true |
Write incident files |
logshark.incidents.flush.seconds |
60 |
How often new incidents are written |
logshark.incidents.retention.days |
90 |
Days of incident files kept |
logshark.incidents.seed.days |
7 |
Days reloaded at startup |
Set these in MetaMinerServer.properties only on advice from Infolytik Support. See Configuration Files.