All Integrations
CloudGCP Service Account + Cloud Monitoring API

Google Spanner Integration

Transaction latency, read/write throughput, and node CPU utilization monitoring for Spanner. Detect CPU hotspots and contention before they impact your globally distributed transactions.

Setup

How It Works

01

Create a GCP Service Account

Create a service account with the Monitoring Viewer and Cloud Spanner Viewer roles. TigerOps uses read-only access to collect Spanner instance and database metrics.

02

Enable Cloud Monitoring API

Enable the Cloud Monitoring API and Cloud Spanner API in your GCP project. TigerOps will pull transaction latency, CPU utilization, and throughput metrics from Cloud Monitoring.

03

Configure TigerOps Spanner

Enter your project credentials and select the Spanner instances and databases to monitor. TigerOps auto-discovers all instances and databases across your project.

04

Set Latency and CPU Alerts

Define transaction latency SLOs, node CPU utilization thresholds, and storage utilization alerts. TigerOps fires predictive alerts before CPU saturation impacts transaction throughput.

Capabilities

What You Get Out of the Box

Transaction Latency Tracking

Monitor read and read-write transaction latency at P50, P99, and P999 percentiles per database. TigerOps correlates latency spikes with CPU hot spots, lock contention, and query plan changes.

Read/Write Throughput

Track read operation counts, write operation counts, and bytes transferred per Spanner instance. TigerOps alerts when throughput approaches node capacity limits before performance degrades.

Node CPU Utilization

Monitor CPU utilization per Spanner node and identify CPU hot spots causing transaction latency. TigerOps recommends node count increases when CPU smoothed utilization approaches the 65% recommended threshold.

Storage Utilization

Track storage used per database and instance. TigerOps monitors storage growth rate and projects when you will reach the per-node storage limit, giving advance warning for capacity planning.

Lock & Contention Metrics

Monitor lock wait times and transaction abort rates. High abort rates indicate write contention hotspots. TigerOps surfaces the tables and key ranges experiencing the most contention.

Multi-Region Replica Health

For multi-region Spanner configurations, TigerOps monitors replica replication lag, leader election events, and per-region request distribution to ensure global consistency SLAs are met.

Configuration

Spanner Integration Setup

Connect your Cloud Spanner instances to TigerOps with GCP service account credentials.

tigerops-spanner.yaml
# TigerOps Cloud Spanner Integration
# Required IAM roles:
#   roles/monitoring.viewer
#   roles/spanner.viewer

integrations:
  gcp_spanner:
    project_id: "your-gcp-project-id"
    credentials_file: "./tigerops-sa-key.json"

    # Instances to monitor (empty = all instances)
    instances:
      - prod-main-instance
      - prod-analytics-instance

    scrape_interval: 60s

    metrics:
      - spanner.googleapis.com/instance/cpu/utilization
      - spanner.googleapis.com/instance/cpu/utilization_by_priority
      - spanner.googleapis.com/instance/storage/used_bytes
      - spanner.googleapis.com/instance/session_count
      - spanner.googleapis.com/api/api_request_count
      - spanner.googleapis.com/api/request_latencies

    alerts:
      # Regional: alert at 65%, multi-region: 45%
      cpu_utilization_regional_percent: 65
      cpu_utilization_multiregion_percent: 45
      storage_utilization_percent: 70
      transaction_abort_rate_per_second: 100
      read_latency_p99_ms: 200
      write_latency_p99_ms: 500
FAQ

Common Questions

What is the recommended CPU utilization threshold for Spanner that TigerOps uses?

Google recommends keeping Cloud Spanner CPU utilization below 65% for regional instances and below 45% for the high-priority CPU of multi-region instances. TigerOps applies these defaults as alert thresholds and notifies you when they are approached.

Can TigerOps monitor Spanner at the database level rather than just the instance level?

Yes. TigerOps collects metrics at both the instance level (CPU, storage) and the database level (operation counts, latency, transaction rates). You can filter and drill down to individual databases in the TigerOps Spanner dashboard.

How does TigerOps detect transaction hotspots in Spanner?

TigerOps monitors the spanner.googleapis.com/instance/cpu/utilization_by_priority metric alongside transaction abort rates. When CPU hotspots correlate with high abort rates, TigerOps flags the hotspot condition and recommends key range analysis via the Spanner Query Insights API.

Does TigerOps support Spanner fine-grained access control (FGAC)?

TigerOps uses a service account with the roles/spanner.viewer role for metric collection via the Cloud Monitoring API. It does not require access to table-level data and is compatible with databases using fine-grained access control.

Can TigerOps help with Spanner capacity planning for node count decisions?

Yes. TigerOps tracks CPU utilization trends over time and projects when you will need additional nodes based on current growth rates. It also monitors throughput utilization to help you understand whether CPU or throughput will be the binding constraint.

Get Started

Keep Your Spanner Transactions Fast and Predictable

CPU hotspot detection, transaction latency SLOs, and capacity planning for Cloud Spanner. Connect in minutes.