RankDisco Troubleshooting Guide

This guide covers common issues, error messages, and step-by-step resolution procedures for RankDisco operations.

Common Issues
Classification Failures
Queue Issues
Workflow Failures
API Errors
Performance Issues
Cost Overruns
Debug Endpoints
Logging
Escalation Path

Common Issues

1. Domain Onboarding Stuck in "Pending" State

Symptoms:

Domain shows onboard_status = 'pending' for extended period
No backlinks or keywords appearing

Error Messages:

[domain-onboard] DRAIN_MODE active, delaying messages
[domain-onboard] Workflow failed, falling back to queue

Resolution:

Check if DRAIN_MODE is enabled:

curl https://<worker>/api/admin/queues/status

Verify workflow binding exists in wrangler.toml:

[[workflows]]
name = "domain-onboard"
binding = "DOMAIN_ONBOARD_WORKFLOW"

Force restart the onboarding:

curl -X POST https://<worker>/api/admin/domains/onboard \
  -H "Content-Type: application/json" \
  -d '{"domain": "example.com"}'

Check workflow status:

curl https://<worker>/api/admin/workflow/domain-onboard/<workflow_id>

2. URLs Not Classifying

Symptoms:

URLs have page_type = NULL
classification_source = 'skipped_domain_unavailable'

Error Messages:

[url-classify] URL not found in database: https://... - will retry
[url-classify] Giving up on classify_url after 3 attempts

Resolution:

Check for race conditions (URL not yet in DB when classification queued):

curl -X POST https://<worker>/api/admin/domains/queue-classification \
  -H "Content-Type: application/json" \
  -d '{"domain": "example.com", "limit": 100}'

Force reclassification:

curl -X POST https://<worker>/api/admin/domains/reclassify-all \
  -H "Content-Type: application/json" \
  -d '{"domain": "example.com", "page_type": null, "limit": 100}'

Check classification failures:

curl https://<worker>/api/admin/domains/classification-failures

3. ClickHouse Connection Failures

Symptoms:

Harvests complete but data not appearing in analytics
Error logs showing connection timeouts

Error Messages:

ClickHouse connection failed: ECONNREFUSED
ClickHouse authentication failed

Resolution:

Verify ClickHouse connectivity:
```
curl https://<worker>/test/clickhouse
```

Check credentials in Wrangler secrets:

wrangler secret list
# Verify: CLICKHOUSE_HOST, CLICKHOUSE_USER, CLICKHOUSE_PASSWORD, CLICKHOUSE_DATABASE

Test direct connection:

curl "https://<CLICKHOUSE_HOST>:8443/?user=<USER>&password=<PASS>&database=<DB>&query=SELECT%201"

4. DataForSEO Rate Limits

Symptoms:

Backlink/keyword fetches failing
Costs spiking unexpectedly

Error Messages:

DataForSEO rate limit exceeded
DataForSEO API error: 429

Resolution:

Check current quotas:
```
curl https://<worker>/api/admin/costs
```

Adjust rate limits in Wrangler vars:

[vars]
DATAFORSEO_LABS_LIMIT = "100"
DATAFORSEO_LABS_MAX_REQUESTS = "50"

Reset budget tracking if stuck:

wrangler kv:key delete --binding=DFS_BUDGETS "daily_budget"

5. Apple iTunes Rate Limiting

Symptoms:

App details failing to fetch
itms-appss:// URLs appearing in logs

Error Messages:

[app-details] rate limited by iTunes API
iTunes returned 403/429 status

Resolution:

Verify ZenRows API key is set:
```
wrangler secret list | grep ZENROWS
```

Check ZenRows quota:

curl https://<worker>/api/admin/costs/zenrows

The worker automatically backs off - wait 30-60 minutes for recovery.

6. Cloudflare Images Upload Failures

Symptoms:

Icons not uploading for apps/domains
Original URLs used as fallback

Error Messages:

CF Images upload failed: 401 Unauthorized
CF Images: missing API token

Resolution:

Verify CF Images secrets:

wrangler secret list | grep CF_IMAGES
# Required: CF_IMAGES_ACCOUNT_ID, CF_IMAGES_API_TOKEN, CF_IMAGES_ACCOUNT_HASH

Test the Images API directly:

curl -X POST "https://api.cloudflare.com/client/v4/accounts/<ACCOUNT_ID>/images/v1" \
  -H "Authorization: Bearer <API_TOKEN>" \
  -F "url=https://example.com/icon.png"

7. Vectorize Index Empty or Stale

Symptoms:

Classification always falling through to LLM
High LLM costs for common domains

Error Messages:

[vectorize] No matches found with score > 0.5
Vectorize index empty or unavailable

Resolution:

Check Vectorize stats:

curl https://<worker>/api/admin/classifier/stats

Initialize seed examples:

curl -X POST https://<worker>/api/admin/classifier/init-vectorize

Bulk load from classified domains:

curl -X POST https://<worker>/api/admin/classifier/bulk-load-vectorize?min_confidence=0.85&limit=1000

8. Workflow Timeout Errors

Symptoms:

Workflows stuck in "running" state
Operations timing out after 2 minutes

Error Messages:

Workflow still running after timeout
Workflow errored: Step execution exceeded timeout

Resolution:

Check workflow status:

curl https://<worker>/api/admin/workflow/<workflow-type>/<workflow-id>

Workflows have built-in retry logic. Wait for automatic recovery.
For persistent failures, check if external APIs are down:
- DataForSEO status page
- Cloudflare status page

9. D1 Database Errors

Symptoms:

Insert/update failures
"Too many subrequests" errors

Error Messages:

D1_ERROR: too many subrequests
D1_ERROR: UNIQUE constraint failed
D1_ERROR: SQLITE_BUSY

Resolution:

Too many subrequests: Batch operations are limited to ~50 statements per batch. Check batch sizes in code.

UNIQUE constraint: Check for duplicate data:

SELECT url_hash, COUNT(*) FROM urls GROUP BY url_hash HAVING COUNT(*) > 1;

SQLITE_BUSY: Reduce concurrent writes or implement backoff.

10. Missing Environment Variables

Symptoms:

Various "not configured" errors
Features silently disabled

Error Messages:

DATAFORSEO_LOGIN not configured
AI binding not configured
VECTORIZE_BACKLINK binding required

Resolution:

Check all required bindings in wrangler.toml:

[vars]
# Required vars listed here

[[kv_namespaces]]
binding = "LOOKUP_CACHE"

[[d1_databases]]
binding = "DB"

[[queues.producers]]
binding = "DOMAIN_ONBOARD_QUEUE"

[[vectorize]]
binding = "VECTORIZE_BACKLINK"

[ai]
binding = "AI"

Verify secrets are set:
```
wrangler secret list
```

Classification Failures

Why URLs Fail to Classify

Classification uses a multi-stage pipeline. Failures can occur at any stage:

Stage	Cost	Common Failures
0. Cache	FREE	Cache miss (expected)
1. Rules	FREE	No matching pattern
2. Vectorize	$0.0001	Index empty, low similarity
3. Content	$0.001-0.01	Fetch blocked, timeout
4. LLM	$0.001	Rate limit, malformed response

Debugging Classification

Test single URL classification:

curl -X POST https://<worker>/api/admin/classifier/classify \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com/article", "options": {"force_llm": false}}'

Test rules only (no external calls):

curl -X POST https://<worker>/api/admin/classifier/classify-rules \
  -H "Content-Type: application/json" \
  -d '{"urls": ["https://example.com/blog/post"]}'

Debug content parsing:

curl -X POST https://<worker>/api/admin/classifier/debug-content \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

Test LLM classification:

curl -X POST https://<worker>/api/admin/classifier/classify-llm \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'

Classification Source Values

Source	Meaning
`rules`	Matched by rules engine (FREE)
`rules_vectorize`	Rules + Vectorize similarity
`llm`	LLM was used for classification
`cache`	Retrieved from cache
`human_feedback`	Manually corrected
`failed`	Classification failed permanently
`skipped_domain_unavailable`	Domain fetch failed

Low Confidence Classifications

Classifications below 65% confidence are flagged for review.

View low confidence URLs:

curl https://<worker>/api/admin/classifier/url-samples?max_confidence=65

Get classification stats:

curl https://<worker>/api/admin/classification-stats

Submit correction:

curl -X POST https://<worker>/api/admin/classifier/corrections \
  -H "Content-Type: application/json" \
  -d '{
    "url_id": 123,
    "is_correct": false,
    "corrected_page_type": "blog_post"
  }'

Queue Issues

Queue Status Monitoring

curl https://<worker>/api/admin/queues/status

Response interpretation:

{
  "summary": {
    "active_queues": 4,
    "total_messages": 150,
    "processing": 25,
    "failed": 3,
    "rate_per_min": 45
  },
  "queues": [
    {
      "name": "url-classify",
      "messages": 100,
      "status": "processing"
    }
  ]
}

Messages Stuck in Queue

Symptoms:

Queue message count not decreasing
status: "idle" with messages pending

Resolution:

Check if DRAIN_MODE is enabled:

# Check
curl https://<worker>/api/admin/config | grep DRAIN_MODE

# Disable
wrangler kv:key delete --binding=LOOKUP_CACHE "DRAIN_MODE"

Verify queue consumers are bound:

[[queues.consumers]]
queue = "url-classify"
max_batch_size = 50
max_batch_timeout = 30

Check for worker errors in logs:
```
wrangler tail --format pretty
```

Dead Letter Queue (DLQ) Overflow

Messages that fail 3 times are moved to DLQ (if configured).

Resolution:

Check DLQ contents via Cloudflare Dashboard
Common causes of repeated failures:
- Invalid URLs that can't be parsed
- Domains that permanently block fetches
- Database constraint violations

Process failed messages manually or purge:

# Via Cloudflare Dashboard: Queues > [queue-name] > Dead Letter Queue

Retry Exhaustion

The system uses exponential backoff with max 3 retries:

Retry 1: 5 seconds delay
Retry 2: 10 seconds delay
Retry 3: 20 seconds delay
After 3 failures: Message ACKed, logged to classification_failures

Check failed classifications:

curl https://<worker>/api/admin/domains/classification-failures

Workflow Failures

Common Workflow Errors

Error	Cause	Resolution
`Workflow errored: Step execution exceeded timeout`	External API slow	Increase timeout or retry
`Failed to resolve domain_id`	Domain not in DB	Ensure domain exists first
`Workflow still running after timeout`	Long-running operation	Wait or check manually
`DOMAIN_ONBOARD_WORKFLOW not configured`	Missing binding	Add to wrangler.toml

Workflow Recovery

Workflows are durable and automatically resume after failures.

Check workflow status:

curl https://<worker>/api/admin/workflow/domain-onboard/<id>

View workflow history:

curl https://<worker>/api/admin/workflow/domain-onboard/<id>/history

Manually trigger retry:

curl -X POST https://<worker>/api/admin/domains/onboard \
  -d '{"domain": "example.com"}'

DomainOnboardWorkflow Failures

Step failures and remediation:

Step	Failure	Resolution
`init-state`	DB write failed	Check D1 connectivity
`ensure-domain`	Hash collision	Rare, retry usually works
`parallel-fetch-all`	DataForSEO timeout	API may be overloaded
`store-backlinks`	Batch too large	Reduce batch size
`store-keywords`	Constraint violation	Check for duplicates
`finalize`	Status update failed	Retry or manual update

DomainClassifyWorkflow Failures

Stage progression:

Stage 1 (TLD Rules) -> Stage 2 (Known Domains) -> Stage 3 (Vectorize) -> 
Stage 4 (Homepage) -> Stage 5 (LLM)

If workflow stops early, check which stage failed:

curl https://<worker>/api/admin/classifier/domain?domain=example.com

API Errors

Error Codes Reference

Status	Error	Meaning	Resolution
400	`domain is required`	Missing required parameter	Include domain in request
400	`AI binding not configured`	Missing wrangler.toml config	Add AI binding
400	`VECTORIZE_BACKLINK binding required`	Missing Vectorize	Add Vectorize binding
401	`Unauthorized`	Missing/invalid auth	Check API keys
404	`Domain not found`	Domain not in database	Onboard domain first
404	`URL not found`	URL not in database	Check URL_ID validity
429	`Rate limit exceeded`	Too many requests	Implement backoff
500	`D1_ERROR`	Database error	Check D1 logs
500	`Workflow failed`	Internal workflow error	Check workflow status

Common API Error Responses

Missing credentials:

{
  "error": "DataForSEO credentials not configured",
  "hint": "Set DATAFORSEO_LOGIN and DATAFORSEO_PASSWORD in wrangler.toml"
}

Invalid URL:

{
  "error": "url is required",
  "status": 400
}

Workflow timeout:

{
  "warning": "Workflow still running after timeout",
  "workflow_id": "abc123",
  "status": "running",
  "status_url": "/api/admin/workflow/domain-classify/abc123"
}

Performance Issues

Slow Classification Responses

Symptoms:

Classification taking > 30 seconds
Timeouts on batch operations

Diagnosis:

Check which stage is slow:

curl -X POST https://<worker>/api/admin/classifier/classify \
  -d '{"url": "...", "use_workflow": true}'
# Response includes stages_run with timing

Check external API latency:
- DataForSEO Instant Pages: typically 1-3s
- Workers AI: typically 0.5-2s
- Vectorize: typically 50-200ms

Resolution:

Skip expensive stages for bulk operations:

curl -X POST https://<worker>/api/admin/classifier/classify-rules \
  -d '{"urls": [...]}'

Use async mode for batch classification:

curl -X POST https://<worker>/api/admin/classifier/classify \
  -d '{"url": "...", "async": true}'

High Latency on Queue Processing

Symptoms:

Queue backlog growing
Processing rate < message arrival rate

Resolution:

Increase batch size (up to 100):

[[queues.consumers]]
max_batch_size = 100

Reduce per-message processing time:
- Enable skip_content: true when SERP data available
- Use rules-only classification for low-priority URLs
Scale horizontally (multiple consumers):
- Cloudflare automatically scales queue consumers

Timeout Handling

Default timeouts:

Worker request: 30 seconds
Workflow step: 2-10 minutes (configurable)
Queue message visibility: 5 minutes

Handling timeouts:

// In workflow
await step.do("my-step", {
  timeout: "5 minutes",
  retries: { limit: 2, delay: "30 seconds" }
}, async () => { ... });

Cost Overruns

Monitoring Costs

Real-time cost tracking:

curl https://<worker>/api/admin/costs

Daily breakdown:

curl https://<worker>/api/admin/costs/daily?days=7

By service:

curl https://<worker>/api/admin/costs/summary

Cost Per Operation

Service	Operation	Cost
DataForSEO	Backlinks fetch	~$0.04/1000
DataForSEO	Keywords fetch	~$0.03/1000
DataForSEO	Domain summary	~$0.02/call
DataForSEO	Instant Pages	~$0.000125/page
Workers AI	LLM (8B model)	~$0.0001/call
Workers AI	LLM (70B model)	~$0.0002/call
Workers AI	Embeddings	~$0.00001/call
ZenRows	Basic credit	~$0.000276/credit
ZenRows	Premium credit	~$0.0069/credit

Budget Exceeded Handling

Symptoms:

Operations silently skipped
budget_exceeded: true in logs

Resolution:

Check current budget status:

curl https://<worker>/api/admin/costs/summary

Reset daily budget (emergency):

wrangler kv:key delete --binding=DFS_BUDGETS "daily_budget_$(date +%Y-%m-%d)"

Increase budget limits:

[vars]
DAILY_COST_LIMIT_USD = "50.00"
DATAFORSEO_DAILY_BUDGET = "25.00"

Reducing Costs

Use rules-only classification first:
- Rules engine is FREE
- Only escalate to LLM when needed
Leverage SERP data for ranking URLs:
- Avoids $0.001+ content fetch per URL
- Uses title/description from DataForSEO
Enable Vectorize learning:
- High-confidence classifications teach the system
- Reduces LLM calls over time

Set appropriate limits:

{
  "backlinks_limit": 100,
  "keywords_limit": 100
}

Debug Endpoints

Classification Debugging

Endpoint	Method	Purpose
`/api/admin/classifier/classify`	POST	Classify single URL (full pipeline)
`/api/admin/classifier/classify-rules`	POST	Rules engine only (FREE)
`/api/admin/classifier/classify-llm`	POST	LLM stage only
`/api/admin/classifier/debug-content`	POST	Debug content fetch
`/api/admin/classifier/parse-content`	POST	Parse page content
`/api/admin/classifier/test`	GET	Test with sample URLs
`/api/admin/classifier/stats`	GET	Index and seed stats

Domain Debugging

Endpoint	Method	Purpose
`/api/admin/domains/status?domain=X`	GET	Check onboarding status
`/api/admin/domains/list`	GET	List domains with filters
`/api/admin/domains/classification-failures`	GET	View failed classifications
`/api/admin/domains/recent-onboards`	GET	Recent onboarding history
`/api/admin/domains/properties?domain_id=X`	GET	Get brand properties

Queue Debugging

Endpoint	Method	Purpose
`/api/admin/queues/status`	GET	All queue statuses
`/api/admin/domains/queue-classification`	POST	Queue URLs for classification
`/api/admin/domains/reclassify-all`	POST	Reclassify URLs matching criteria

Cost Debugging

Endpoint	Method	Purpose
`/api/admin/costs`	GET	Overall cost stats
`/api/admin/costs/daily`	GET	Daily breakdown
`/api/admin/costs/recent`	GET	Recent cost entries
`/api/admin/costs/zenrows`	GET	ZenRows quota status
`/api/admin/costs/summary`	GET	Comprehensive summary

Workflow Debugging

Endpoint	Method	Purpose
`/api/admin/workflow/<type>/<id>`	GET	Workflow status
`/api/admin/workflow/<type>/<id>/history`	GET	Workflow step history

Logging

Viewing Logs

Real-time logs:

wrangler tail --format pretty

Filtered by tag:

wrangler tail --format pretty --search "url-classify"

Log Prefixes

Prefix	Component
`[url-classify]`	URL classification consumer
`[domain-classify]`	Domain classification consumer
`[domain-onboard]`	Domain onboarding consumer
`[admin-*]`	Admin endpoints
`[workflow]`	Workflow execution
`[cache]`	Cache operations
`[dataforseo]`	DataForSEO API calls

What to Look For

Successful classification:

[url-classify] Classified URL example.com: page_type=blog_post, channel=media (85% confidence, $0.000100)

Failed classification:

[url-classify] Error processing classify_url, retrying (2 left): URL not found in database
[url-classify] Giving up on classify_url after 3 attempts: timeout

Queue processing:

[url-classify] Batch complete: 45/50 succeeded (PARALLEL), 5 target URLs, cost: $0.004500

Workflow progress:

[domain-onboard] Starting onboarding for domain: example.com (id: 123)
[domain-onboard] Fetched 100 backlinks
[domain-onboard] Stored 95 keywords
[domain-onboard] Finalized with status: completed

Log Levels

info: Normal operations
warn: Non-fatal issues (cache misses, retries)
error: Failures requiring attention

Escalation Path

Level 1: Self-Service (0-15 minutes)

Check this troubleshooting guide
Verify configuration in wrangler.toml
Check Cloudflare Dashboard for queue/worker status
Review recent logs with wrangler tail

Level 2: Investigation (15-60 minutes)

Use debug endpoints to isolate issue
Check external service status pages:
- Cloudflare Status
- DataForSEO Status
Review recent deployments that may have introduced issues
Check D1/KV/R2 metrics in Cloudflare Dashboard

Level 3: Engineering Escalation

When to escalate:

Data loss or corruption suspected
Security incident
Persistent failures after Level 1-2 resolution attempts
Cost overrun > 5x normal daily spend

Information to include:

Timestamp of when issue started
Affected domains/URLs
Error messages from logs
Steps already attempted
Impact assessment (# of affected customers/operations)

Escalation contacts:

Primary: On-call engineer (check PagerDuty/Slack)
Secondary: Engineering lead
Data issues: Data engineering team

Recovery Procedures

Full system restart:

# Redeploy worker
wrangler deploy

# Clear all caches (use with caution)
wrangler kv:bulk delete --binding=LOOKUP_CACHE

# Reset queue (via Cloudflare Dashboard)

Selective recovery:

# Reclassify specific domain's URLs
curl -X POST https://<worker>/api/admin/domains/reclassify-all \
  -d '{"domain": "example.com", "limit": 500}'

# Re-onboard specific domain
curl -X POST https://<worker>/api/admin/domains/onboard \
  -d '{"domain": "example.com"}'

Quick Reference Card

Health Checks

# Overall system health
curl https://<worker>/test/clickhouse

# Queue health
curl https://<worker>/api/admin/queues/status

# Cost status
curl https://<worker>/api/admin/costs/summary

# Classification stats
curl https://<worker>/api/admin/classification-stats

Emergency Commands

# Enable drain mode (stops all queue processing)
wrangler kv:key put --binding=LOOKUP_CACHE "DRAIN_MODE" "true"

# Disable drain mode
wrangler kv:key delete --binding=LOOKUP_CACHE "DRAIN_MODE"

# Force redeploy
wrangler deploy --force

Useful Queries

-- Find stuck onboardings
SELECT domain, onboard_status, onboard_started_at 
FROM domains 
WHERE onboard_status = 'pending' 
  AND onboard_started_at < datetime('now', '-1 hour');

-- Find classification failures
SELECT domain, COUNT(*) as failed_count 
FROM classification_failures 
GROUP BY domain 
ORDER BY failed_count DESC 
LIMIT 10;

-- Find unclassified URLs
SELECT domain, COUNT(*) as unclassified 
FROM urls 
WHERE page_type IS NULL 
GROUP BY domain 
ORDER BY unclassified DESC 
LIMIT 20;

Table of Contents​

Common Issues​

1. Domain Onboarding Stuck in "Pending" State​

2. URLs Not Classifying​

3. ClickHouse Connection Failures​

4. DataForSEO Rate Limits​

5. Apple iTunes Rate Limiting​

6. Cloudflare Images Upload Failures​

7. Vectorize Index Empty or Stale​

8. Workflow Timeout Errors​

9. D1 Database Errors​

10. Missing Environment Variables​

Classification Failures​

Why URLs Fail to Classify​

Debugging Classification​

Classification Source Values​

Low Confidence Classifications​

Queue Issues​

Queue Status Monitoring​

Messages Stuck in Queue​

Dead Letter Queue (DLQ) Overflow​

Retry Exhaustion​

Workflow Failures​

Common Workflow Errors​

Workflow Recovery​

DomainOnboardWorkflow Failures​

DomainClassifyWorkflow Failures​

API Errors​

Error Codes Reference​

Common API Error Responses​

Performance Issues​

Slow Classification Responses​

High Latency on Queue Processing​

Timeout Handling​

Cost Overruns​

Monitoring Costs​

Cost Per Operation​

Budget Exceeded Handling​

Reducing Costs​

Debug Endpoints​

Classification Debugging​

Domain Debugging​

Queue Debugging​

Cost Debugging​

Workflow Debugging​

Logging​

Viewing Logs​

Log Prefixes​

What to Look For​

Log Levels​

Escalation Path​

Level 1: Self-Service (0-15 minutes)​

Level 2: Investigation (15-60 minutes)​

Level 3: Engineering Escalation​

Recovery Procedures​

Quick Reference Card​

Health Checks​

Emergency Commands​

Useful Queries​

Table of Contents

Common Issues

1. Domain Onboarding Stuck in "Pending" State

2. URLs Not Classifying

3. ClickHouse Connection Failures

4. DataForSEO Rate Limits

5. Apple iTunes Rate Limiting

6. Cloudflare Images Upload Failures

7. Vectorize Index Empty or Stale

8. Workflow Timeout Errors

9. D1 Database Errors

10. Missing Environment Variables

Classification Failures

Why URLs Fail to Classify

Debugging Classification

Classification Source Values

Low Confidence Classifications

Queue Issues

Queue Status Monitoring

Messages Stuck in Queue

Dead Letter Queue (DLQ) Overflow

Retry Exhaustion

Workflow Failures

Common Workflow Errors

Workflow Recovery

DomainOnboardWorkflow Failures

DomainClassifyWorkflow Failures

API Errors

Error Codes Reference

Common API Error Responses

Performance Issues

Slow Classification Responses

High Latency on Queue Processing

Timeout Handling

Cost Overruns

Monitoring Costs

Cost Per Operation

Budget Exceeded Handling

Reducing Costs

Debug Endpoints

Classification Debugging

Domain Debugging

Queue Debugging

Cost Debugging

Workflow Debugging

Logging

Viewing Logs

Log Prefixes

What to Look For

Log Levels

Escalation Path

Level 1: Self-Service (0-15 minutes)

Level 2: Investigation (15-60 minutes)

Level 3: Engineering Escalation

Recovery Procedures

Quick Reference Card

Health Checks

Emergency Commands

Useful Queries