Handle possible nil in profiling config

[adrien-f]

Hello!

I've been bitten by this panic:

goroutine 276 [running]:
github.com/parca-dev/parca/pkg/scrape.LabelsByProfiles(0x4007237d58?, 0x39bdb00?)
	github.com/parca-dev/parca/pkg/scrape/target.go:233 +0x120
github.com/parca-dev/parca/pkg/scrape.targetsFromGroup(0x400f5b6ea0, 0x400187c420, {0x0?, 0x46e?, 0x500?}, 0x4007237f10)
	github.com/parca-dev/parca/pkg/scrape/target.go:370 +0x158
github.com/parca-dev/parca/pkg/scrape.(*scrapePool).Sync(0x400ffa0b40, {0x40078ba000, 0x46e, 0x40073a8e40?})
	github.com/parca-dev/parca/pkg/scrape/scrape.go:218 +0x228
github.com/parca-dev/parca/pkg/scrape.(*Manager).reload.func1(0x0?, {0x40078ba000?, 0x4001880a90?, 0x4000fa8160?})
	github.com/parca-dev/parca/pkg/scrape/manager.go:217 +0x28
created by github.com/parca-dev/parca/pkg/scrape.(*Manager).reload in goroutine 62
	github.com/parca-dev/parca/pkg/scrape/manager.go:216 +0x1c4

while having a config file with a custom pprof:

scrape_configs:
      - job_name: kubernetes-nodejs
        scrape_interval: 60s
        scrape_timeout: 70s
        profiling_config:
          pprof_config:
            allocs:
              enabled: false
            block:
              enabled: false
            cpu:
              path: /api/debug/pprof/profile
            memory:
              enabled: false
            mutex:
              enabled: false
      ...

I think Go will short-circuit the condition and not evaluate the right expression right?

[CLAassistant]

All committers have signed the CLA.

[adrien-f]

Hello @metalmatze , do you know what's happening with the CI?

  /usr/bin/docker pull quay.io/containers/podman:v4.7.2@sha256:15a3734034751edd5615f6324e10029c04089bdcb9d161c35e9fd20b08aa9748
  Error response from daemon: manifest for quay.io/containers/podman@sha256:15a3734034751edd5615f6324e10029c04089bdcb9d161c35e9fd20b08aa9748 not found: manifest unknown: manifest unknown
  Error: Docker pull failed with exit code 1

Anything I need to pull/rebase?

[brancz]

Apologies for this, unfortunately the podman team keeps deleting container images all the time. Can you rebase this again? It was just updated on main a couple of hours ago.

[brancz]

Thank you!!