UNAVAILABLE - Unavailable: Service Temporarily Down

#Common Causes

• →Service Maintenance: GCP performs planned maintenance on service infrastructure. Maintenance is typically scheduled and announced. This is transient—waiting for maintenance to complete and retrying helps.
• →Service Overload: The service is experiencing high load and can't process requests. The service may throttle or temporarily reject requests. This is transient—retrying with exponential backoff helps.
• →Regional Outage: The service in a specific region experiences an unplanned outage. Outages may affect all resources in that region. This is transient—waiting for service restoration and retrying helps, or using a different region may work.
• →Service-Specific Behavior: Availability behavior varies by service. Some services recover quickly, others may take longer. Timing isn't guaranteed.

✓Solutions

1. 1Step 1: Diagnose - Check GCP status page for known issues: Visit https://status.cloud.google.com/ and check for service outages or maintenance affecting your service/region.
2. 2Step 2: Diagnose - Check if the error is consistent by retrying the operation multiple times: gcloud compute instances list --project PROJECT_ID If it consistently fails, it may be a persistent issue.
3. 3Step 3: Diagnose - Check if the error is region-specific by trying operations in a different region: gcloud compute zones list --format="table(name,status)" Then retry your operation in a different zone/region.
4. 4Step 4: Fix - Implement retry logic with exponential backoff. Retry the operation with increasing delays (e.g., 1s, 2s, 4s, 8s).
5. 5Step 5: Fix - For regional outages, try operations in a different region if possible: gcloud compute instances create INSTANCE_NAME --zone DIFFERENT_ZONE --project PROJECT_ID
6. 6Step 6: Verify - If retries succeed, the issue was transient. If retries consistently fail, check the GCP status page and consider contacting GCP support.

</>Code Examples

1# This script implements retry logic with exponential backoff for UNAVAILABLE errors
2
3PROJECT_ID="my-project"
4MAX_RETRIES=5
5RETRY_DELAY=1
6
7# Retry function with exponential backoff
8retry_with_backoff() {
9  local max_attempts=$1
10  local delay=$2
11  local attempt=1
12  shift 2
13  local command="$@"
14  
15  while [ $attempt -le $max_attempts ]; do
16    echo "Attempt $attempt of $max_attempts..."
17    
18    if eval "$command"; then
19      echo "Success! Operation completed on attempt $attempt"
20      return 0
21    fi
22    
23    if [ $attempt -lt $max_attempts ]; then
24      echo "Operation failed. Waiting $delay seconds before retry..."
25      sleep $delay
26      delay=$((delay * 2))  # Exponential backoff
27      attempt=$((attempt + 1))
28    else
29      echo "Failed after $max_attempts attempts"
30      echo "Please check:"
31      echo "  1. GCP Status page: https://status.cloud.google.com/"
32      echo "  2. Try a different zone: gcloud compute zones list"
33      echo "  3. Contact GCP support if the issue persists"
34      return 1
35    fi
36  done
37}
38
39# Usage: Retry a gcloud command
40retry_with_backoff $MAX_RETRIES $RETRY_DELAY gcloud compute instances list --project $PROJECT_ID

↗Related Errors