GET /sla-metrics

Returns real-time service quality (SLA) metrics: percentile latencies, success/error rates, and service availability.

Endpoint

GET https://places.gotravelyzer.com/sla-metrics

Authentication

This endpoint does not require authentication and is not counted against your quotas.

Parameters

No parameters required.

Request

curl "https://places.gotravelyzer.com/sla-metrics"

Response

Success (200 OK)

{
  "period": "last_24h",
  "window": "5m",
  "latency": {
    "p50": 3,
    "p95": 18,
    "p99": 35
  },
  "requests": {
    "total": 124500,
    "successRate": 99.8,
    "errorRate": 0.2
  },
  "availability": {
    "uptime": 99.99
  }
}

Response Fields

Metadata

Field	Type	Description
`period`	`string`	Observation period (e.g., `last_24h`)
`window`	`string`	Aggregation window (e.g., `5m` = 5 minutes)

Latency Metrics

Field	Type	Description
`latency.p50`	`number`	Median latency in ms (50% of requests are faster)
`latency.p95`	`number`	95th percentile in ms
`latency.p99`	`number`	99th percentile in ms

Latency values are clamped to a minimum of 1ms for better readability.

Request Metrics

Field	Type	Description
`requests.total`	`number`	Total number of requests processed
`requests.successRate`	`number`	Success rate (0-100%)
`requests.errorRate`	`number`	Error rate (0-100%)

Availability

Field	Type	Description
`availability.uptime`	`number`	Availability percentage (0-100%)

HTTP Codes

Code	Status	Description
`200`	OK	Metrics returned successfully
`503`	Service Unavailable	Metrics service unavailable

Use Cases

Health Monitoring

Check that metrics are within acceptable thresholds:

#!/bin/bash
 
METRICS=$(curl -s "https://places.gotravelyzer.com/sla-metrics")
 
P95=$(echo $METRICS | jq '.latency.p95')
SUCCESS_RATE=$(echo $METRICS | jq '.requests.successRate')
UPTIME=$(echo $METRICS | jq '.availability.uptime')
 
echo "Latency P95: ${P95}ms"
echo "Success rate: ${SUCCESS_RATE}%"
echo "Availability: ${UPTIME}%"
 
# Check thresholds
if [ $(echo "$P95 > 200" | bc) -eq 1 ]; then
  echo "ALERT: High P95 latency!"
fi
 
if [ $(echo "$SUCCESS_RATE < 99" | bc) -eq 1 ]; then
  echo "ALERT: Low success rate!"
fi

Supervision Dashboard

Retrieve metrics for a dashboard:

# Retrieve formatted metrics
curl -s "https://places.gotravelyzer.com/sla-metrics" | jq '{
  latency_p50: .latency.p50,
  latency_p95: .latency.p95,
  latency_p99: .latency.p99,
  success_rate: .requests.successRate,
  error_rate: .requests.errorRate,
  uptime: .availability.uptime,
  total_requests: .requests.total
}'

Automatic Alerts

Metrics-based alert script:

#!/bin/bash
 
ALERT_P95_THRESHOLD=200      # ms
ALERT_SUCCESS_THRESHOLD=99   # %
ALERT_UPTIME_THRESHOLD=99.5  # %
 
METRICS=$(curl -s "https://places.gotravelyzer.com/sla-metrics")
 
P95=$(echo $METRICS | jq '.latency.p95')
SUCCESS=$(echo $METRICS | jq '.requests.successRate')
UPTIME=$(echo $METRICS | jq '.availability.uptime')
 
ALERTS=""
 
if [ $(echo "$P95 > $ALERT_P95_THRESHOLD" | bc) -eq 1 ]; then
  ALERTS+="High P95 latency: ${P95}ms (threshold: ${ALERT_P95_THRESHOLD}ms)\n"
fi
 
if [ $(echo "$SUCCESS < $ALERT_SUCCESS_THRESHOLD" | bc) -eq 1 ]; then
  ALERTS+="Low success rate: ${SUCCESS}% (threshold: ${ALERT_SUCCESS_THRESHOLD}%)\n"
fi
 
if [ $(echo "$UPTIME < $ALERT_UPTIME_THRESHOLD" | bc) -eq 1 ]; then
  ALERTS+="Low availability: ${UPTIME}% (threshold: ${ALERT_UPTIME_THRESHOLD}%)\n"
fi
 
if [ -n "$ALERTS" ]; then
  echo -e "ALERTS DETECTED:\n$ALERTS"
  # Send notification (Slack, email, etc.)
else
  echo "All metrics are within normal thresholds"
fi

Interpreting Metrics

Latency (P50, P95, P99)

Percentile	Meaning	Typical Target
P50	Median experience	< 50ms
P95	Experience of 95% of users	< 200ms
P99	Common worst case	< 500ms

Success Rate

Value	Interpretation
> 99.9%	Excellent
99-99.9%	Normal
95-99%	Degraded
< 95%	Critical problem

Availability (Uptime)

Value	Equivalent Downtime (monthly)
99.99%	~4 minutes
99.9%	~43 minutes
99%	~7 hours

Technical Notes

Metrics are calculated over a rolling 5-minute window
Latency values are in milliseconds
Rates are expressed as percentage (0-100)
This endpoint is not subject to rate limits
These are informational metrics, not contractual guarantees

Next Steps

SLA History

Performance history over 24-48h.

Health Check

Check if the API is operational.

Coverage

Geographic coverage by country.