Promtails cannot push messages to Loki (or sent some messages)

The logs of Loki are not shown or partly shown in Grafana.

My environment is Docker Desktop with Kubernetes running on it. I used the Loki stack to install it. I can connect via either http://loki:3100 or via http://loki.grafana-loki.svc.cluster.local.

  1. In Grafana, the datasource config of Loki gives frequently an error like

Failed to load log volume for this query
parse error at line 1, col 67: syntax error: unexpected IDENTIFIER

  1. In the Promtail logs I can see:

2024-10-03 19:36:11 level=warn ts=2024-10-03T17:36:11.959668723Z caller=logging.go:86 traceID=549a9dd21956c96a orgID=fake msg="POST /loki/api/v1/push (500) 254.011µs Response: "at least 1 live replicas required, could only find 0 - unhealthy instances: 10.1.2.214:9095\n" ws: false; Content-Length: 410; Content-Type: application/x-protobuf; User-Agent: promtail/2.9.3; "
2024-10-03 19:36:12 level=error ts=2024-10-03T17:36:12.478371512Z caller=compactor.go:374 msg=“error asking ring for who should run the compactor, will check again” err=“at least 1 healthy replica required, could only find 0 - unhealthy instances: 10.1.2.214:9095”

In the Promtail dashboard I can see that logs are scraped.

The Kubernetes I see the following pods run:

NAME READY STATUS RESTARTS AGE
loki-0 1/1 Running 0 97s
loki-grafana-7c95854c5c-sx2g2 2/2 Running 0 97s
loki-promtail-5pfnc 1/1 Running 0 97s

I can Grafana datasource Loki I see via the label “namespace” only the “grafana-loki”, “kube-system” and “argocd” namespaces, not the default.

When selecting the default namespace, which is not in the list, then I get only the error of the unexpected IDENTIFIER and no logs at all.

I frequently get the message in Grafana “Failed to load log volume for this query
parse error at line 1, col 73: syntax error: unexpected IDENTIFIER”

I can see some logs.