r3kzi / clamav-prometheus-exporter Goto Github PK

Hi,

I've deployed clamav recently and started scraping from prometheus.
There is an issue i found that mem gauges are returning "NaN" instead memory usage.

Has anyone experienced such a thing?

Example:

# HELP clamav_build_info Shows ClamAV Build Info
# TYPE clamav_build_info gauge
clamav_build_info{clamav_version="0.104.2",database_version="26532"} 1
# HELP clamav_mem_heap_bytes Shows heap memory usage in bytes
# TYPE clamav_mem_heap_bytes gauge
clamav_mem_heap_bytes NaN
# HELP clamav_mem_mmap_bytes Shows mmap memory usage in bytes
# TYPE clamav_mem_mmap_bytes gauge
clamav_mem_mmap_bytes NaN
# HELP clamav_mem_used_bytes Shows used memory in bytes
# TYPE clamav_mem_used_bytes gauge
clamav_mem_used_bytes NaN

Hi, I was doing some testing running as a sidecar to a pod running clamav and a http interface, the http interface is using the same TCP port to send scan requests, when I ran some tests the exporter would exit causing its container to restart. Interested if there is a way we could handle this.

Apparently clamav-prometheus-exporter does not react properly to kill -SIGTERM

So far with old containerd < v1.5 this did not matter, as containerd did not react properly on reboot when used in docker v20.10.x with live-restore. (containerd/containerd#5502)

Now that this is fixed, clamav-prometheus-exporter is the last container hanging.

When I enter the container and send kill -SIGTERM nothing happens.

Can you try and clarify, if that is something I can fix on my side?

Thank You!

Hi,
would it be possible to push a new image to dockerhub and/or configure automated builds if it's available to you please?

For now I've rebuilt a docker image manually on my end which I use :)

that would be great to provide build in the release section for those who don'thave the possibility to use docker

We connect to our clamd via sockets, and in order to use your exporter, we'd need to turn on TCP listening. Since an exporter typically runs on the same machine as the thing that it is exporting, it might be a useful change to provide a socket connection capability to speak to clamd as an alternative to the TCP based one.

Using current head of the branch, when I try to run the exporter, I get this error:

 panic: runtime error: index out of range [2] with length 1
 goroutine 26 [running]:
 github.com/r3kzi/clamav-prometheus-exporter/pkg/collector.(*Collector).Collect(0xc0001ae100, 0xc000055f60?)
         /usr/pkg/mod/github.com/r3kzi/[email protected]/pkg/collector/collector.go:139 +0x1452
 github.com/prometheus/client_golang/prometheus.(*Registry).Gather.func1()
         /usr/pkg/mod/github.com/prometheus/[email protected]/prometheus/registry.go:444 +0xfb
 created by github.com/prometheus/client_golang/prometheus.(*Registry).Gather
         /usr/pkg/mod/github.com/prometheus/[email protected]/prometheus/registry.go:455 +0x4c5

When I try to install, I'm getting the following:

GOPATH=/usr go get github.com/r3kzi/clamav-prometheus-exporter
# github.com/prometheus/procfs/internal/util
usr/src/github.com/prometheus/procfs/internal/util/parse.go:69:15: undefined: os.ReadFile
usr/src/github.com/prometheus/procfs/internal/util/parse.go:78:15: undefined: os.ReadFile
usr/src/github.com/prometheus/procfs/internal/util/readfile.go:36:9: undefined: io.ReadAll

Im trying to connect it like sidecar in my deployment with clamav, but unfortunately it seems not working, vmservicescrape dont see metrics from exporter :(

Thanks for making this! I was doing this the harder way.

Do you have a grafana dashboard that you use for these values? Would be nice if you did and it was included in the examples directory!

I would suggest changing
clamav_database_age 13580.878680335
to a full timestamp in seconds
clamav_database_age 1677590624770.1213
as it is a value that doesn't change over time and has the same content.
Users can calculate the delta at query time easily enough.

Hi there! First off thanks so much for this exporter. I am hoping to put in a request to add a metric for Infected Files so that I may create an alert around it. I am currently sending an email from a shell script if an infected file is found on an instance, however I believe a more robust solution would be to implement the monitor and alert within Prometheus/Grafana. I have looked at the collector.go file and wonder if my idea and request is a possibility?