You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
By default, `startupProbe` is enabled and `readinessProbe` and `livenessProbe` are disabled. You can enable/disable them via `.startupProbe.enabled``.readinessProbe.enabled``.livenessProbe.enabled` in respective node type.
385
390
@@ -411,6 +416,22 @@ edgeNode:
411
416
periodSeconds: 5
412
417
```
413
418
419
+
#### Distributor Probes
420
+
421
+
By default, `startupProbe`, `readinessProbe` and `livenessProbe` are enabled for this component in both full distributed and Hub-Nodes mode.
422
+
423
+
There is a script in chart `configs/distributor/distributorProbe.sh` is loaded into ConfigMap and mounted to the container is used by `livenessProbe`. You can customize the script via `--set-file distributorConfigMap.extraScripts.distributorProbe\.sh=/path/to/your_script.sh` or set via YAML values.
424
+
425
+
There are some reports on a scenario that would be difficult to reproduce or rare: `Grid UI is accessible but no nodes can be fetched or registered. Or something like there are few requests in session queue but could not be accepted. After restarting the Distributor, the issue is resolved`. Based on that, a proactive approach to do automatic restart whenever detecting it is not healthy via `livenessProbe` and the condition check is executed. The script queries GraphQL endpoint to get `sessionCount`, and `sessionQueueSize`. If the `sessionQueueSize` is greater than 0 and `sessionCount` is 0 until the `failureThreshold`, the Distributor will be restarted. You can adjust the threshold as well as interval via probe settings.
426
+
427
+
#### Router Probes
428
+
429
+
By default, `startupProbe`, `readinessProbe` and `livenessProbe` are enabled for this component in full distributed mode.
430
+
431
+
There is a script in chart `configs/router/routerProbe.sh` loaded into ConfigMap and mounted to the container is used by `livenessProbe`. You can customize the script via `--set-file routerConfigMap.extraScripts.routerProbe\.sh=/path/to/your_script.sh` or set via YAML values.
432
+
433
+
The script checks GraphQL endpoint is reachable. If the `http_code` is not `200` until the `failureThreshold`, the Router will be restarted. You can adjust the threshold as well as interval via probe settings.
434
+
414
435
### Configuration extra scripts mount to container
415
436
416
437
This is supported for containers of browser node, video recorder and video uploader. By default, in these containers, there are scripts, config files implemented. In case you want to customize or replace them with your own implementation. Instead of forking the chart, use volume mount. Now, from your external files, you can insert them into ConfigMap via Helm CLI `--set-file` or compose them in your own YAML values file and pass to Helm CLI `--values` when deploying chart. Any files name that you defined will be picked up into ConfigMap and mounted to the container.
echo"$(date ${ts_format}) DEBUG [${probe_name}] - Could not construct GraphQL endpoint, it can be set directly via SE_GRID_GRAPHQL_URL. Bypass the probe checks for now."
echo"$(date ${ts_format}) DEBUG [${probe_name}] - Could not construct GraphQL endpoint, it can be set directly via SE_GRID_GRAPHQL_URL. Bypass the probe checks for now."
0 commit comments