sentry-error-triage（哨兵错误分类）

Sentry Integration

Use sentry-cli and the Sentry REST API to monitor errors, triage issues, manage releases, and upload source maps.

Setup

# Install
npm i -g @sentry/cli
# — or —
pip install sentry-cli

# Auth (set once, used by all commands)
export SENTRY_AUTH_TOKEN="sntrys_..."   # Settings → Auth Tokens → Create
export SENTRY_ORG="my-org"
export SENTRY_PROJECT="my-project"

# Verify
sentry-cli info

Generate a token at sentry.io → Settings → Auth Tokens with scopes: project:read, project:releases, org:read, event:read.

CLI Commands

Releases

# Create a release (version from git)
sentry-cli releases new "$(sentry-cli releases propose-version)"

# Set commits (auto-detect from git)
sentry-cli releases set-commits "$VERSION" --auto

# Finalize (marks release as deployed)
sentry-cli releases finalize "$VERSION"

# Create + finalize in one step
sentry-cli releases new "$VERSION" --finalize

# Record a deploy
sentry-cli deploys new -r "$VERSION" -e production

# List releases
sentry-cli releases list

Source Maps

# Upload source maps for a release
sentry-cli sourcemaps upload ./dist --release "$VERSION"

# With URL prefix (match hosted paths)
sentry-cli sourcemaps upload ./dist --release "$VERSION" --url-prefix "~/static/js"

# Validate before upload
sentry-cli sourcemaps explain --release "$VERSION" --org "$SENTRY_ORG" --project "$SENTRY_PROJECT"

Send Test Event

sentry-cli send-event -m "Test event from CLI"
sentry-cli send-event -m "Deploy check" -t environment:production -t release:1.0.0

Monitor (Cron Monitoring)

# Wrap a command — Sentry tracks if it runs and succeeds
sentry-cli monitors run <monitor-slug> -- <command>
sentry-cli monitors run backup-job -- ./run-backup.sh

REST API (for queries the CLI doesn't cover)

Base URL: https://sentry.io/api/0 Auth header: Authorization: Bearer $SENTRY_AUTH_TOKEN

List Issues

# All unresolved issues for a project
curl -s "https://sentry.io/api/0/projects/$SENTRY_ORG/$SENTRY_PROJECT/issues/?query=is:unresolved" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" | jq '.[].title'

# Organization-wide issues (sorted by last seen)
curl -s "https://sentry.io/api/0/organizations/$SENTRY_ORG/issues/?query=is:unresolved&sort=date" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" | jq '.[] | {id, title, count, lastSeen}'

# Filter by level, time, assignment
curl -s "https://sentry.io/api/0/organizations/$SENTRY_ORG/issues/?query=is:unresolved+level:error+lastSeen:>2d" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN"

Get Issue Details + Events

# Issue details
curl -s "https://sentry.io/api/0/issues/$ISSUE_ID/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" | jq '{title, status, count, firstSeen, lastSeen}'

# Latest events for an issue (stack traces, breadcrumbs)
curl -s "https://sentry.io/api/0/issues/$ISSUE_ID/events/?full=true" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" | jq '.[0].entries'

Resolve / Ignore Issues

# Resolve
curl -s "https://sentry.io/api/0/issues/$ISSUE_ID/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -X PUT -H "Content-Type: application/json" \
  -d '{"status": "resolved"}'

# Resolve in next release
curl -s "https://sentry.io/api/0/issues/$ISSUE_ID/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -X PUT -H "Content-Type: application/json" \
  -d '{"status": "resolvedInNextRelease"}'

# Ignore for 24 hours
curl -s "https://sentry.io/api/0/issues/$ISSUE_ID/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -X PUT -H "Content-Type: application/json" \
  -d '{"status": "ignored", "statusDetails": {"ignoreDuration": 1440}}'

# Assign to team member
curl -s "https://sentry.io/api/0/issues/$ISSUE_ID/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -X PUT -H "Content-Type: application/json" \
  -d '{"assignedTo": "user@example.com"}'

Bulk Resolve

# Resolve multiple issues at once
curl -s "https://sentry.io/api/0/projects/$SENTRY_ORG/$SENTRY_PROJECT/issues/" \
  -H "Authorization: Bearer $SENTRY_AUTH_TOKEN" \
  -X PUT -H "Content-Type: application/json" \
  -d '{"id": ["123","456","789"], "status": "resolved"}'

Triage Workflow

When asked to check or triage Sentry errors:

1. List unresolved issues sorted by frequency: query=is:unresolved&sort=freq
2. For the top issues, fetch latest event with full stack trace
3. Analyze the stack trace — identify the failing function, file, and line
4. Check if the error is in code the agent can access (read the file, suggest a fix)
5. Classify: critical (data loss, crash), high (user-facing errors), medium (degraded experience), low (cosmetic, logs)
6. Resolve issues that have confirmed fixes deployed; ignore transient errors

Notes

• Self-hosted Sentry: replace sentry.io with your instance URL
• Rate limits: 40 requests/min for free tier, respect Retry-After headers
• The CLI respects .sentryclirc files for project-level config
• Use --log-level debug on any CLI command for troubleshooting