Couchbase Integration
Monitor bucket metrics, view query latency, XDCR replication lag, and N1QL performance across your Couchbase deployments. Get AI-powered DCP backlog detection and rebalance impact analysis.
How It Works
Connect via Couchbase REST API
TigerOps connects to the Couchbase Management REST API (port 8091) using a read-only monitoring user. Cluster, bucket, node, and XDCR metrics are pulled via the /pools/default and /pools/default/buckets endpoints.
Enable Prometheus Metrics (CB 7.x)
For Couchbase Server 7.x, enable the native Prometheus metrics endpoint. TigerOps scrapes /metrics on each node for per-service stats including Data, Query, Index, Search, Analytics, and Eventing service metrics.
Configure XDCR Monitoring
TigerOps monitors XDCR replication streams — tracking replication lag, data replicated bytes, and error rates per stream and remote cluster. Cross-datacenter lag alerts fire before replication falls behind your SLO.
Set Bucket and Query Alerts
Define thresholds for bucket memory saturation, disk ejection rates, N1QL query latency, and XDCR replication lag. TigerOps correlates DCP backlog growth with vBucket rebalance and failover events.
What You Get Out of the Box
Bucket Metrics & Memory
Per-bucket resident ratio, memory used vs quota, item counts, get/set hit rates, cache miss ratios, and disk ejection rates. Alert when bucket memory saturation causes increased disk read latency.
View Query Latency
Per-design-document view query execution time, rows returned per query, and stale vs non-stale query ratios. TigerOps surfaces views with high read amplification or missing index coverage.
XDCR Replication Lag
Per-stream XDCR lag (in milliseconds and bytes), replication throughput, error counts, and filtered document rates. TigerOps alerts when cross-datacenter replication lag exceeds your DR SLO.
N1QL Query Performance
N1QL service request rates, p50/p95/p99 execution latency, request errors, and index advisor recommendations. Identify which N1QL queries are performing full bucket scans vs using primary indexes.
Node & Service Health
Per-node CPU, memory, disk usage, and service-level health status for Data, Index, Query, Search, and Analytics services. TigerOps detects vBucket rebalance events and failover operations in real time.
AI DCP Backlog Detection
TigerOps AI monitors DCP backlog growth rates across buckets and fires early alerts when DCP consumer lag indicates that indexing, XDCR, or eventing consumers are falling behind — before data loss risk materializes.
TigerOps Agent Config for Couchbase
Connect TigerOps to Couchbase Server via the REST API and native Prometheus endpoint (CB 7.x).
# TigerOps Couchbase integration config
# Place at /etc/tigerops/conf.d/couchbase.yaml
integrations:
- name: couchbase
type: couchbase
config:
# Couchbase cluster management endpoint
management_url: https://couchbase.internal:18091
username: tigerops_monitor
password: "${COUCHBASE_MONITOR_PASSWORD}"
tls:
enabled: true
ca_cert: /etc/ssl/couchbase/ca.pem
# Couchbase Server version (auto-detected if not set)
# server_version: "7.2"
# Buckets to monitor (empty = all buckets)
buckets:
- default
- sessions
- products
- analytics
# Metrics to collect
metrics:
bucket_stats: true # Memory, ops, cache, disk
node_stats: true # Per-node CPU, memory, disk
xdcr_stats: true # Replication lag, throughput
query_service: true # N1QL request rates, latency
index_service: true # GSI index scan rates, memory
search_service: true # FTS query latency, item counts
eventing_service: true # Function execution, DCP lag
# Native Prometheus scrape (Couchbase 7.x only)
prometheus:
enabled: true
port: 8093 # query service Prometheus port
# Per-node scraping via pod IPs (Kubernetes)
kubernetes:
enabled: true
namespace: couchbase
label_selector: "app=couchbase"
scrape_interval: 30s
remote_write:
endpoint: https://ingest.atatus.net/api/v1/write
bearer_token: "${TIGEROPS_API_KEY}"
alerts:
bucketResidentRatioMin: 0.85
xdcrLagMs: 5000
n1qlP99LatencyMs: 200
dcpBacklogItems: 100000Common Questions
Which Couchbase versions does TigerOps support?
TigerOps supports Couchbase Server 6.x, 7.x, and Couchbase Capella (fully managed cloud). For Server 7.x, TigerOps uses the native Prometheus endpoint for richer per-service metrics. For Server 6.x, TigerOps uses the REST API with automatic version detection.
What permissions does the Couchbase monitoring user need?
The monitoring user needs the Cluster Admin RBAC role with read-only access, or a custom role with cluster_admin[read], bucket_full_access[read] for each monitored bucket, and replication_admin[read] for XDCR metrics. No data read or write permissions are required.
How does TigerOps monitor XDCR lag without a remote cluster connection?
TigerOps collects XDCR stats from the source cluster REST API at /pools/default/buckets/{bucket}/stats with xdcr_* keys. This gives replication lag and throughput without requiring connectivity to the remote cluster. Remote cluster health is tracked separately via the /pools/default/remoteClusters endpoint.
Can TigerOps detect vBucket rebalance operations?
Yes. TigerOps monitors the /pools/default/rebalanceProgress endpoint and fires alerts when a rebalance begins, including the estimated duration and percentage complete. Rebalance events are overlaid on memory, disk I/O, and query latency charts for impact analysis.
Does TigerOps support Couchbase Eventing and Analytics services?
Yes. For Couchbase 7.x with Prometheus metrics enabled, TigerOps collects Eventing service function execution rates, timeout counts, and DCP lag, alongside Analytics (CBAS) query latency and active ingestion rates. All services appear in unified cluster health dashboards.
Get Complete Visibility Into Your Couchbase Cluster
Bucket health, XDCR replication lag, N1QL performance, and AI DCP backlog detection. Deploy in 5 minutes.