Skip to content

Cache GitHub Actions job metadata in logs output #59039

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
pelikhan merged 4 commits into from
Sep 6, 2026
Merged

Cache GitHub Actions job metadata in logs output #59039

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
8e7ba1c
Cache GitHub jobs API responses
Copilot Sep 6, 2026
f2c1e1a
Secure cached jobs API data
Copilot Sep 6, 2026
2fa1775
Harden jobs cache persistence
Copilot Sep 6, 2026
96d1cd6
Refine jobs cache error handling
Copilot Sep 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions pkg/cli/audit_analysis_fanout.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@ func launchCoreAuditAnalyses(g *errgroup.Group, gctx context.Context, results *a
expName, expVariant, _ := firstExperimentAssignment(extractExperimentData(runOutputDir))

launchMetricsAnalysis(g, gctx, results, runOutputDir, verbose, run.WorkflowPath)
launchJobDetailsAnalysis(g, gctx, results, run.DatabaseID, verbose)
launchJobDetailsAnalysis(g, gctx, results, run.DatabaseID, runOutputDir, verbose)
runAuditAnalysis(g, gctx, verbose, "extractMissingToolsFromRun", "Failed to extract missing tools", func(v []MissingToolReport) {
results.missingTools = v
}, func() ([]MissingToolReport, error) {
Expand Down Expand Up @@ -100,12 +100,12 @@ func launchMetricsAnalysis(g *errgroup.Group, gctx context.Context, results *aud
}

// launchJobDetailsAnalysis exclusively writes results.jobDetails and results.failedJobCount.
func launchJobDetailsAnalysis(g *errgroup.Group, gctx context.Context, results *auditAnalysisResults, runID int64, verbose bool) {
func launchJobDetailsAnalysis(g *errgroup.Group, gctx context.Context, results *auditAnalysisResults, runID int64, runOutputDir string, verbose bool) {
g.Go(func() error {
if err := gctx.Err(); err != nil {
return err
}
jobDetails, failedJobCount, err := fetchJobDetailsWithCounts(gctx, runID, verbose)
jobDetails, failedJobCount, err := fetchJobDetailsWithCounts(gctx, runID, runOutputDir, verbose)
if err != nil {
if gctx.Err() != nil {
return gctx.Err()
Expand Down
68 changes: 37 additions & 31 deletions pkg/cli/logs_github_api.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,12 +16,14 @@ import (
"fmt"
"os"
"os/exec"
"path/filepath"
"slices"
"strconv"
"strings"
"time"

"github.com/github/gh-aw/pkg/console"
"github.com/github/gh-aw/pkg/constants"
"github.com/github/gh-aw/pkg/logger"
"github.com/github/gh-aw/pkg/workflow"
)
Expand Down Expand Up @@ -76,49 +78,53 @@ func buildCreatedFilter(startDate, endDate, beforeDate string) string {
// call and returns the full detail slice together with the count of failed jobs.
// It is the single source of truth for the jobs endpoint; fetchJobDetails and
// fetchJobStatuses are thin wrappers that each return only the value they need.
func fetchJobDetailsWithCounts(ctx context.Context, runID int64, verbose bool) ([]JobInfoWithDuration, int, error) {
func fetchJobDetailsWithCounts(ctx context.Context, runID int64, outputDir string, verbose bool) ([]JobInfoWithDuration, int, error) {
logsGitHubAPILog.Printf("Fetching job details: runID=%d", runID)
if verbose {
fmt.Fprintln(os.Stderr, console.FormatVerboseMessage(fmt.Sprintf("Fetching job details for run %d", runID)))
}

output, err := workflow.RunGHCombinedContext(ctx, "Fetching job details...", "api",
fmt.Sprintf("repos/{owner}/{repo}/actions/runs/%d/jobs", runID),
"--jq", ".jobs[] | {name: .name, status: .status, conclusion: (.conclusion // \"\"), started_at: .started_at, completed_at: .completed_at, steps: ((.steps // []) | map({name: .name, status: .status, conclusion: (.conclusion // \"\")}))}")
fmt.Sprintf("repos/{owner}/{repo}/actions/runs/%d/jobs?per_page=100", runID),
"--paginate", "--slurp")
if err != nil {
if verbose {
fmt.Fprintln(os.Stderr, console.FormatVerboseMessage(fmt.Sprintf("Failed to fetch job details for run %d: %v", runID, err)))
}
return nil, 0, err
}

var jobs []JobInfoWithDuration
failedJobs := 0
lines := strings.SplitSeq(strings.TrimSpace(string(output)), "\n")
for line := range lines {
if strings.TrimSpace(line) == "" {
continue
}

var job JobInfo
if err := json.Unmarshal([]byte(line), &job); err != nil {
if verbose {
fmt.Fprintln(os.Stderr, console.FormatVerboseMessage("Failed to parse job info: "+line))
}
continue
if outputDir != "" {
responsePath := filepath.Join(outputDir, jobsAPIResponseFileName)
if err := os.WriteFile(responsePath, output, constants.FilePermSensitive); err != nil {

@github-actions github-actions Bot Sep 6, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[/tdd] No test covers os.WriteFile failing (e.g. outputDir not existing, permission denied) — that path currently returns a hard error from fetchJobDetailsWithCounts, which callers treat as "no job data available" even though this is a caching side-effect, not a fetch failure.

💡 Suggested test 
func TestFetchJobDetailsWithCounts_CacheWriteFailure(t *testing.T) {
    // outputDir points at a non-existent parent (or a read-only dir)
    // assert jobs/failedJobs are still returned, or that the error is
    // clearly scoped to the cache-write concern.
}

Worth deciding intentionally whether a cache-write failure should fail job detail retrieval at all, since every caller already discards fetchJobDetailsWithCounts errors as non-fatal (fetchJobDetails, fetchJobStatuses, launchJobDetailsAnalysis) — so today it silently degrades, which may be fine, but should be covered by a test rather than implicit.

@copilot please address this.

return nil, 0, fmt.Errorf("failed to cache jobs API response: %w", err)
}
logsGitHubAPILog.Printf("Cached jobs API response: path=%s", responsePath)
}

jobWithDuration := JobInfoWithDuration{JobInfo: job}
if !job.StartedAt.IsZero() && !job.CompletedAt.IsZero() {
jobWithDuration.Duration = job.CompletedAt.Sub(job.StartedAt)
}
jobs = append(jobs, jobWithDuration)
var responses []struct {
Jobs []JobInfo `json:"jobs"`
}
if err := json.Unmarshal(output, &responses); err != nil {

@github-actions github-actions Bot Sep 6, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

[/diagnosing-bugs] Resilience regression: parsing now aborts entirely on any malformed byte in the whole paginated response, whereas the old per-line loop skipped just the bad job and kept the rest.

💡 Details

The previous implementation iterated line-by-line and logged+skipped any job that failed to unmarshal, so a single corrupted/partial page didn't discard every other job. Now json.Unmarshal(output, &responses) is all-or-nothing — one bad page (e.g. truncated network response, unexpected field type) zeroes out job info for the entire run, even though most pages were fine. Since this directly affects the failed-job count and audit summaries, consider decoding per-page (or per-job) with a fallback that logs and continues, matching the old resilience behavior. Add a test with one malformed page among several valid ones to lock in the desired behavior.

@copilot please address this.

return nil, 0, fmt.Errorf("failed to parse jobs API response: %w", err)
}

if isFailureConclusion(job.Conclusion) {
failedJobs++
logsGitHubAPILog.Printf("Found failed job: name=%s, conclusion=%s", job.Name, job.Conclusion)
if verbose {
fmt.Fprintln(os.Stderr, console.FormatVerboseMessage(fmt.Sprintf("Found failed job '%s' with conclusion '%s'", job.Name, job.Conclusion)))
var jobs []JobInfoWithDuration
failedJobs := 0
for _, response := range responses {
for _, job := range response.Jobs {
jobWithDuration := JobInfoWithDuration{JobInfo: job}
if !job.StartedAt.IsZero() && !job.CompletedAt.IsZero() {
jobWithDuration.Duration = job.CompletedAt.Sub(job.StartedAt)
}
jobs = append(jobs, jobWithDuration)

if isFailureConclusion(job.Conclusion) {
failedJobs++
logsGitHubAPILog.Printf("Found failed job: name=%s, conclusion=%s", job.Name, job.Conclusion)
if verbose {
fmt.Fprintln(os.Stderr, console.FormatVerboseMessage(fmt.Sprintf("Found failed job '%s' with conclusion '%s'", job.Name, job.Conclusion)))
}
}
}
}
Expand All @@ -130,8 +136,8 @@ func fetchJobDetailsWithCounts(ctx context.Context, runID int64, verbose bool) (
// fetchJobDetails gets detailed job information including durations for a workflow run.
// Errors from the underlying API call are suppressed so that callers can continue
// processing even when job data is unavailable (e.g. missing permissions).
func fetchJobDetails(ctx context.Context, runID int64, verbose bool) ([]JobInfoWithDuration, error) {
jobs, _, err := fetchJobDetailsWithCounts(ctx, runID, verbose)
func fetchJobDetails(ctx context.Context, runID int64, outputDir string, verbose bool) ([]JobInfoWithDuration, error) {
jobs, _, err := fetchJobDetailsWithCounts(ctx, runID, outputDir, verbose)
if err != nil {
// Don't fail the entire operation if we can't get job info
return nil, nil
Expand All @@ -143,7 +149,7 @@ func fetchJobDetails(ctx context.Context, runID int64, verbose bool) ([]JobInfoW
// Errors from the underlying API call are suppressed so that callers can continue
// processing even when job data is unavailable (e.g. missing permissions).
func fetchJobStatuses(ctx context.Context, runID int64, verbose bool) (int, error) {
_, failedJobs, err := fetchJobDetailsWithCounts(ctx, runID, verbose)
_, failedJobs, err := fetchJobDetailsWithCounts(ctx, runID, "", verbose)
if err != nil {
// Don't fail the entire operation if we can't get job info
return 0, nil
Expand Down Expand Up @@ -189,7 +195,7 @@ type ListWorkflowRunsOptions struct {
// not the total number of matching runs the user wants to find.
//
// The processedCount and targetCount parameters are used to display progress in the spinner message.
func listWorkflowRunsWithPagination(opts ListWorkflowRunsOptions) ([]WorkflowRun, int, error) {
func listWorkflowRunsWithPagination(opts ListWorkflowRunsOptions) ([]WorkflowRun, int, error) { //nolint:largefunc // Existing run listing keeps pagination, error classification, and filtering together.
logsGitHubAPILog.Printf("Listing workflow runs: workflow=%s, limit=%d, startDate=%s, endDate=%s, ref=%s", opts.WorkflowName, opts.Limit, opts.StartDate, opts.EndDate, opts.Ref)
args := []string{"run", "list", "--json", "databaseId,number,url,status,conclusion,workflowName,createdAt,startedAt,updatedAt,event,headBranch,headSha,displayTitle"}

Expand Down
26 changes: 20 additions & 6 deletions pkg/cli/logs_github_api_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -229,30 +229,44 @@ func TestListWorkflowRunsErrorHandling(t *testing.T) {

func TestFetchJobDetailsWithCountsIncludesSteps(t *testing.T) {
fakeBinDir := testutil.TempDir(t, "fake-gh-*")
outputDir := t.TempDir()
fakeGH := filepath.Join(fakeBinDir, "gh")
argsLogPath := filepath.Join(fakeBinDir, "gh-args.log")
fakeGHScript := "#!/bin/sh\n" +
"printf '%s\\n' \"$*\" >> \"" + argsLogPath + "\"\n" +
"cat <<'EOF'\n" +
"{\"name\":\"agent\",\"status\":\"completed\",\"conclusion\":\"failure\",\"started_at\":\"2026-06-28T01:31:00Z\",\"completed_at\":\"2026-06-28T01:33:00Z\",\"steps\":[{\"name\":\"Set up job\",\"status\":\"completed\",\"conclusion\":\"success\"},{\"name\":\"Run agent\",\"status\":\"completed\",\"conclusion\":\"failure\"}]}\n" +
"[{\"total_count\":1,\"jobs\":[{\"id\":42,\"run_id\":28307653871,\"run_attempt\":2,\"html_url\":\"https://github.com/github/gh-aw/actions/runs/28307653871/job/42\",\"status\":\"completed\",\"conclusion\":\"failure\",\"created_at\":\"2026-06-28T01:30:00Z\",\"started_at\":\"2026-06-28T01:31:00Z\",\"completed_at\":\"2026-06-28T01:33:00Z\",\"name\":\"agent\",\"runner_name\":\"GitHub Actions 1\",\"steps\":[{\"name\":\"Set up job\",\"status\":\"completed\",\"conclusion\":\"success\",\"number\":1,\"started_at\":\"2026-06-28T01:31:00Z\",\"completed_at\":\"2026-06-28T01:31:10Z\"},{\"name\":\"Run agent\",\"status\":\"completed\",\"conclusion\":\"failure\",\"number\":2,\"started_at\":\"2026-06-28T01:31:10Z\",\"completed_at\":\"2026-06-28T01:33:00Z\"}]}]}]\n" +
"EOF\n"
require.NoError(t, os.WriteFile(fakeGH, []byte(fakeGHScript), 0o755))

t.Setenv("PATH", fakeBinDir+string(os.PathListSeparator)+os.Getenv("PATH"))

jobs, failedJobs, err := fetchJobDetailsWithCounts(context.Background(), 28307653871, false)
jobs, failedJobs, err := fetchJobDetailsWithCounts(context.Background(), 28307653871, outputDir, false)
require.NoError(t, err)
require.Len(t, jobs, 1)
assert.Equal(t, 1, failedJobs, "failed job count should include failed jobs")
assert.Equal(t, 2*time.Minute, jobs[0].Duration, "job duration should still be derived from timestamps")
assert.Equal(t, int64(42), jobs[0].ID, "high-level GitHub job metadata should be preserved")
assert.Equal(t, 2, jobs[0].RunAttempt)
assert.Equal(t, "GitHub Actions 1", jobs[0].RunnerName)
require.Len(t, jobs[0].Steps, 2)
assert.Equal(t, "Run agent", jobs[0].Steps[1].Name, "step names should be parsed from gh api output")
assert.Equal(t, "failure", jobs[0].Steps[1].Conclusion, "step conclusions should be parsed from gh api output")
assert.Equal(t, 2, jobs[0].Steps[1].Number)

argsLog, err := os.ReadFile(argsLogPath)
require.NoError(t, err)
assert.Contains(t, string(argsLog), "repos/{owner}/{repo}/actions/runs/28307653871/jobs", "should query the run jobs API")
assert.Contains(t, string(argsLog), "steps:", "gh jq projection should request step data")
assert.Contains(t, string(argsLog), "repos/{owner}/{repo}/actions/runs/28307653871/jobs?per_page=100", "should query the run jobs API")
assert.Contains(t, string(argsLog), "--paginate --slurp", "should cache all pages of the jobs API response")
assert.NotContains(t, string(argsLog), "--jq", "should cache the complete API response without a projection")

cachedResponse, err := os.ReadFile(filepath.Join(outputDir, jobsAPIResponseFileName))
require.NoError(t, err)
assert.Contains(t, string(cachedResponse), `"total_count":1`)
assert.Contains(t, string(cachedResponse), `"runner_name":"GitHub Actions 1"`)
cachedInfo, err := os.Stat(filepath.Join(outputDir, jobsAPIResponseFileName))
require.NoError(t, err)
assert.Equal(t, os.FileMode(0o600), cachedInfo.Mode().Perm())
}

// TestFetchJobDetailsWithCountsNullConclusion verifies that jobs and steps with null conclusions
Expand All @@ -265,13 +279,13 @@ func TestFetchJobDetailsWithCountsNullConclusion(t *testing.T) {
// A job still in progress has conclusion="" for itself and for any pending steps.
fakeGHScript := "#!/bin/sh\n" +
"cat <<'EOF'\n" +
"{\"name\":\"agent\",\"status\":\"in_progress\",\"conclusion\":\"\",\"started_at\":\"2026-06-28T01:31:00Z\",\"completed_at\":\"0001-01-01T00:00:00Z\",\"steps\":[{\"name\":\"Set up job\",\"status\":\"completed\",\"conclusion\":\"success\"},{\"name\":\"Run agent\",\"status\":\"in_progress\",\"conclusion\":\"\"}]}\n" +
"[{\"total_count\":1,\"jobs\":[{\"name\":\"agent\",\"status\":\"in_progress\",\"conclusion\":null,\"started_at\":\"2026-06-28T01:31:00Z\",\"completed_at\":null,\"steps\":[{\"name\":\"Set up job\",\"status\":\"completed\",\"conclusion\":\"success\"},{\"name\":\"Run agent\",\"status\":\"in_progress\",\"conclusion\":null}]}]}]\n" +
"EOF\n"
require.NoError(t, os.WriteFile(fakeGH, []byte(fakeGHScript), 0o755))

t.Setenv("PATH", fakeBinDir+string(os.PathListSeparator)+os.Getenv("PATH"))

jobs, failedJobs, err := fetchJobDetailsWithCounts(context.Background(), 28307653871, false)
jobs, failedJobs, err := fetchJobDetailsWithCounts(context.Background(), 28307653871, "", false)
require.NoError(t, err)
require.Len(t, jobs, 1, "in-progress jobs with null conclusion should not be dropped")
assert.Equal(t, 0, failedJobs, "in-progress job should not count as failed")
Expand Down
40 changes: 31 additions & 9 deletions pkg/cli/logs_models.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,8 @@ const (
defaultAgentStdioLogPath = "/tmp/gh-aw/agent-stdio.log"
// runSummaryFileName is the name of the summary file created in each run folder
runSummaryFileName = "run_summary.json"
// jobsAPIResponseFileName is the raw GitHub Actions jobs API response cached for each run
jobsAPIResponseFileName = "jobs.json"
// defaultLogsOutputDir is the default directory for downloaded workflow logs
defaultLogsOutputDir = ".github/aw/logs"
)
Expand Down Expand Up @@ -302,19 +304,39 @@ type DownloadResult struct {

// JobInfo represents basic information about a workflow job
type JobInfo struct {
Name string `json:"name"`
Status string `json:"status"`
Conclusion string `json:"conclusion"`
StartedAt time.Time `json:"started_at,omitzero"`
CompletedAt time.Time `json:"completed_at,omitzero"`
Steps []JobStep `json:"steps,omitempty"`
ID int64 `json:"id,omitempty"`
RunID int64 `json:"run_id,omitempty"`
RunURL string `json:"run_url,omitempty"`
RunAttempt int `json:"run_attempt,omitempty"`
NodeID string `json:"node_id,omitempty"`
HeadSha string `json:"head_sha,omitempty"`
URL string `json:"url,omitempty"`
HTMLURL string `json:"html_url,omitempty"`
Status string `json:"status"`
Conclusion string `json:"conclusion"`
CreatedAt time.Time `json:"created_at,omitzero"`
StartedAt time.Time `json:"started_at,omitzero"`
CompletedAt time.Time `json:"completed_at,omitzero"`
Name string `json:"name"`
Steps []JobStep `json:"steps,omitempty"`
CheckRunURL string `json:"check_run_url,omitempty"`
Labels []string `json:"labels,omitempty"`
RunnerID int64 `json:"runner_id,omitempty"`
RunnerName string `json:"runner_name,omitempty"`
RunnerGroupID int64 `json:"runner_group_id,omitempty"`
RunnerGroupName string `json:"runner_group_name,omitempty"`
WorkflowName string `json:"workflow_name,omitempty"`
HeadBranch string `json:"head_branch,omitempty"`
}

// JobStep represents basic information about an individual workflow job step.
type JobStep struct {
Name string `json:"name"`
Status string `json:"status,omitempty"`
Conclusion string `json:"conclusion,omitempty"`
Name string `json:"name"`
Status string `json:"status,omitempty"`
Conclusion string `json:"conclusion,omitempty"`
Number int `json:"number,omitempty"`
StartedAt time.Time `json:"started_at,omitzero"`
CompletedAt time.Time `json:"completed_at,omitzero"`
}

// JobInfoWithDuration extends JobInfo with calculated duration
Expand Down
2 changes: 1 addition & 1 deletion pkg/cli/logs_run_processor.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -540,7 +540,7 @@ func applyRunUsageMetrics(result *DownloadResult, metrics *LogMetrics, runOutput
// RunSummary struct, and writes it to disk. It also sets the agentic-analysis fields on
// result directly so they are available to the caller.
func finalizeAndSaveRunSummary(ctx context.Context, result *DownloadResult, runOutputDir string, metrics LogMetrics, verbose bool) {
jobDetails, jobErr := fetchJobDetails(ctx, result.Run.DatabaseID, verbose)
jobDetails, jobErr := fetchJobDetails(ctx, result.Run.DatabaseID, runOutputDir, verbose)
if jobErr != nil && verbose {
fmt.Fprintln(os.Stderr, console.FormatWarningMessage(fmt.Sprintf("Failed to fetch job details for run %d: %v", result.Run.DatabaseID, jobErr)))
} else {
Expand Down
Loading