Google Cloud Stackdriver

From Luis Gallego Hurtado - Not Another IT guy
Jump to: navigation, search

Stackdriver Logging

Stackdriver Logging receives, indexes and stores log messages from many sources, including Google Cloud Platform, Amazon Web Services, VM instances running stackdriver fluentd agent and user applications.

  • A Stackdrive account can match one or several GCP Projects or AWS accounts.
  • Log Search
  • Log Based Metrics, from different types and apply them to log messages with specific labels.
  • Log Exports, that send them to Cloud Storage (buckets) or Cloud Pub/Sub (log streaming).
  • Log Alerts.
  • Avanced Analytics with BigQuery.
  • Log retention for 30 days or longer with Cloud Storage.


  • Audit logs: admin activity, system events and data access.
  • Agent logs.
  • Logs available in Stackdriver Logging


  • Integration with different GCP products, services and audit logs, Splunk, AWS.
  • All logs are stored into with same structure, a LogEntry, but they can have different payloads.
  • You can have custom logs and perform ingestion via API.
  • AWS Integration / Agent.


Free tier 50GB per month.

Free logs: audit logging and excluded logs

Stackdriver monitoring

Stackdriver Monitoring is a full-stack monitoring tool for Google Cloud Platform, Amazon Web Services and hybrid deployments.

  • A Stackdrive account can match one or several GCP Projects or AWS accounts.
  • Uptime and healthcheck monitoring
  • It supports alerts, custom metrics, existing matrices, custom matrices, time series and labels.
  • You can identify trends and prevent issues.
  • Metric data is deleted after 6 weeks.
  • Time series are deleted when all their points have expired.
  • Data registered from API is shown in few seconds, while CPU utilization is available after 3-4 minutes.
  • Resources are chargeable in stackdriver only when they are running.
  • Pricing: fee per data and API calls. There is a free tier (only GCP, no AWS) with some limitations.


You define resource type, metric, you can filter and group by, and eventually, you can aggregate data by providing an alignement period, an aligner (how different values withing alignement period are computerd) and reducer (how aligned values are grouped).

Stackdriver Debugging

Stackdriver Debugging allows you to inspect your application in Google App Engine and Java, Python and Go.

  • Capture all local and stack variables.
  • Inject logging into a running service by creating LogPoints.
  • Enabled in Googled App Engined by importing and using library.

Stackdriver Error Reporting

Realtime error monitoring and alerting.

  • Stacktrace exploration.
  • Issue tracker integration.
  • Designed for Google App Engined but it can be used with Google Compute Environment and Amazón EC2.
  • Enabled in Googled App Engined by importing and using library.

Stackdriver Trace

Stackdriver Trace is a distributed tracing service that collects latency data for finding performance bottleneck in production apps in Google App Engine, HTTP Load Balancer and applications that use Stackdriver SDK.

  • Latency shift detection.
  • No need to provide credentials in Google Cloud Platform:
    • Enabled in Googled App by default.
    • It can be enabled in Google Kubernetes Engine on creating cluster (by adding trace.append).
    • It must be enabled manually in Google Compute Engine.
  • Outside Google Cloud Platform, credentials must be send by specifying path to key file.
  • Pricing: fee per spans ingested and spans scanned. There is a free tier.