topic Re: Gateways Appear In Skyline Dashboard But No Data in OpenTelemetry/Skyline

Gateways Appear In Skyline Dashboard But No Data

cdav — Wed, 05 Feb 2025 19:03:07 GMT

I have a graf/prom instance running the Skyline dashboards and I have around 10 gateways and 2 managers successfully streaming data and appearing in the grafana dashboard.

I have 2 gateways (a cluster) that appear in the dashboard but with no data.

When looking at the metrics being received in prometheus the majority do not contain entries for the gateways in question e.g "system_memory_usage" or "system_cpu_utilisation" but metrics like "cpview_info" do show entries for the gateways.

I have checked on the gateways the following: cpviewexporter version, otelcol version and cpviewapi version all of which are the same as my working gateways. I have also checked the list of metrics using sklnctl and the default are present like the other gateways.

Can anyone help?

Screenshot of dashboard attached

Re: Gateways Appear In Skyline Dashboard But No Data

Elad_Chomsky — Thu, 06 Feb 2025 08:41:41 GMT

Hi @cdav ,

It might be a problem with CPView - try the following:

- kill -9 $(pidof cpviewd)

- cpview -a off

- cpview -a on

If this is still not working, please open a support ticket to CheckPoint, so we can try to assist you directly.

Re: Gateways Appear In Skyline Dashboard But No Data

cdav — Thu, 06 Feb 2025 10:41:56 GMT

This seems to have made progress - i can now see in the dashboard the installed policy, install timestamp and uptime. The rest remains blank.

I already have a support ticket raised so will see how that progresses.

Re: Gateways Appear In Skyline Dashboard But No Data

Jeremy1 — Wed, 14 May 2025 23:50:51 GMT

Hi team,

I have the same problem on a customer.

Do you have updates from TAC?

Re: Gateways Appear In Skyline Dashboard But No Data

Elad_Chomsky — Thu, 15 May 2025 10:17:39 GMT

Hi @Jeremy1 ,

I suggest to open a ticket to CheckPoint directly, it seems like a problem related to CPview.

Re: Gateways Appear In Skyline Dashboard But No Data

Jeremy1 — Thu, 22 May 2025 15:36:54 GMT

Hi team,

Hope everything is going well.

I was able to solve the problem; in my case were 2 things

1. Different versions of otelcol

I observed that the version of otelcol was different from the most updated version; today it is 116.0; before it was 82.0. The following command can be used to verify the otelcol version:
#/opt/CPotelcol/otelcol --version

These were the error messages I had:

"2025/04/24 09:36:03 failed to upload metrics: context deadline exceeded: rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing: dial unix /opt/CPotelcol/grpc_otlp.sock: connect: no such file or directory"

The 'grpc_otlp.sock' in 82.0 otelcol version suddenly disappeared, so I was having that error.

In 116.0 otelcol version, the .sock appears again, so I was able to send metrics again. It updated automatically, as you can see in the next messages:

admin localhost t +otlp:otelcol:backup:version CPotelcol_0.82.0
Configuration changed from localhost by user admin by the service /bin/dbset
Configuration changed from localhost by user admin by the service /bin/dbset
admin localhost t +otlp:otelcol:active_config_version CPotelcol_0.116.0

2. Issues with mapping metrics.

When I ran the cpview -m command, I noticed that not all the metrics I wanted were sent, so I validated the error messages and showed the following:

"caller=metrics_instruments.go:138 level=warn msg="Metric system.cpu.utilization does not exist in metricsLastSample.Metrics"

It was quite strange since no file had been modified.

For example, if you searched the system.cpu.utilization metric with the following command:
# cpview -m -f M:"system.cpu.utilization" | jq '.'
The output was null

Additionally, I validated that it was found in the metrics repository according to the following command:
#sklnctl otelcol metrics --show

And it was mapped; however, still not sent the metric.

So I decided to stop and start the cpview service with the next commands:

cpview -s off

cpview -a off

and then,

cpview -s on

cpview -a on

Next, I checked if the mapped system.cpu.utilization metric will be found, and in fact, it already appeared:

#cpview -m | jq '.' | grep -i metric\-id

"metric-id": "system.cpu.utilization"

My issue was solved.