How to read verbose logs to track down slow linters?

[trallnag]

I am using the --verbose flag, but I have a hard time seeing how much time individual linters took.

Running [/home/runner/golangci-lint-1.50.1-linux-amd64/golangci-lint run --out-format=github-actions --verbose] in [] ...
  level=info msg="[config_reader] Config search paths: [./ /home/runner/work/kubestat2cloudwatch-aggregator/kubestat2cloudwatch-aggregator /home/runner/work/kubestat2cloudwatch-aggregator /home/runner/work /home/runner /home /]"
  level=info msg="[config_reader] Used config file .golangci.yaml"
  level=info msg="[lintersdb] Active 44 linters: [asasalint asciicheck bidichk bodyclose containedctx contextcheck cyclop decorder dupl durationcheck errcheck errchkjson errname errorlint execinquery exhaustive exportloopref forbidigo gochecknoglobals gochecknoinits gosec gosimple govet ineffassign lll maintidx makezero nestif nilerr nilnil noctx nolintlint nonamedreturns predeclared promlinter staticcheck tagliatelle tenv thelper tparallel typecheck unconvert unused wrapcheck]"
  level=info msg="[loader] Go packages loading at mode 575 (imports|name|deps|files|types_sizes|compiled_files|exports_file) took 11.36667288s"
  level=info msg="[runner/filename_unadjuster] Pre-built 0 adjustments in 2.216713ms"
  level=info msg="[linters_context/goanalysis] analyzers took 3m0.808470226s with top 10 stages: buildir: 1m30.508971326s, buildssa: 1m14.715283445s, inspect: 2.959766473s, nilness: 2.331648051s, exhaustive: 2.185017311s, fact_deprecated: 2.159697523s, printf: 1.783153568s, ctrlflow: 1.554747512s, typedness: 796.881641ms, fact_purity: 709.395884ms"
  level=info msg="[runner] Issues before processing: 1, after processing: 0"
  level=info msg="[runner] Processors filtering stat (out/in): cgo: 1/1, path_prettifier: 1/1, identifier_marker: 1/1, exclude: 1/1, exclude-rules: 0/1, skip_dirs: 1/1, autogenerated_exclude: 1/1, filename_unadjuster: 1/1, skip_files: 1/1"
  level=info msg="[runner] processing took 195.01µs with stages: skip_dirs: 55.703µs, identifier_marker: 42.702µs, autogenerated_exclude: 39.502µs, exclude-rules: 24.402µs, path_prettifier: 23.401µs, cgo: 2.1µs, max_same_issues: 1.3µs, filename_unadjuster: 1.1µs, nolint: 800ns, skip_files: 600ns, uniq_by_line: 500ns, diff: 500ns, severity-rules: 400ns, max_from_linter: 400ns, exclude: 400ns, max_per_file_from_linter: 300ns, path_shortener: 300ns, path_prefixer: 200ns, source_code: 200ns, sort_results: 200ns"
  level=info msg="[runner] linters took 1m1.171524969s with stages: goanalysis_metalinter: 1m1.171226254s"
  level=info msg="File cache stats: 0 entries of total size 0B"
  level=info msg="Memory: 7[27](https://github.com/trallnag/kubestat2cloudwatch-aggregator/actions/runs/3795294726/jobs/6454270911#step:4:29) samples, avg is 7[39](https://github.com/trallnag/kubestat2cloudwatch-aggregator/actions/runs/3795294726/jobs/6454270911#step:4:41).8MB, max is 1140.0MB"
  level=info msg="Execution took 1m12.551027499s"

For example I am seeing that goanalysis_metalinter, buildir, and buildssa took a long time. But are that linters? There are not declared in the config.

Related to #2126.

[ldez]

Hello,

in the core of golangci-lint, there are passes, and linters are just passes.
metalinter is the linter that contains all the linters.
buildir is a pass that creates representations of the elements of Go programs, this pass is used by linters related to staticcheck.
buildssa is a pass that creates representations of the elements of Go programs, this pass is used by several linter that uses a static single-assignment (SSA)

[trallnag]

Thanks for the insight