Monitoring

Stackdriver

• It is the tool provided by GCP for monitoring, logging and debugging
• It gives insides to application health, performance and availability
• Stackdriver Logging lets define metrics based on logs. These metrics can be displayed on dashboards
• Stackdirver Error Reporting: tracks errors in applications and it can notify us when new errors are detected
• Stackdriver Trace: report on application latency and sampling

• Stackdriver Debugger: it connects application production data to the source code for inspecting application state

• Dynamically discovers cloud resources and application services
• We can have deep visibility into our applications in minutes
• Provides us access to powerful data an analytics tools
• Stackdriver offers services for:
  • Monitoring
  • Logging
  • Error Reporting
  • Tracing
  • Debugging
• Supports several third party integrations

Stackdriver Monitoring

• Is at the base os SRE
• It dynamically configures monitoring after resources are deployed
• Allows us to monitor platform, systems an application metrics
• Workspace:
  • Is a root entity that holds monitoring and configuration information in Stackdriver Monitoring
  • We can have as many workspaces as we want, GCP projects can’t be monitored by more than one workspace
  • The first monitored project is the Hosting Project and needs to be specified at the workspace creation
• Stackdriver allows us to create custom dashboards and charts based on the monitored data
• Alerting policies: we can create alerting policies based on monitored data. We can create notifications based on these alerting policies
• Update checks: test the availability of the public services
• Monitoring agent: used to access system resources and application services. It can be installed in compute resources and application services

Stackdriver Logging

• Allows us to store, search, analyze and alert on log data on events from GCP and AWS
• Logging includes storage for logs, an user interface Logs Viewer and an API to manage logs programmatically
• Logs are retained for 30 days, we can export logs into Cloud Storage, BigQuery and Cloud Pub/Sub
• Logging Agent: can be installed on VM instances for gathering logs

Stackdriver Error Reporting

• Counts, analyzes and aggregates errors for running cloud services
• Provides a centralized error management interface

Stackdriver Trace

• It is a distributed tracing system that collects latency data from application systems and generates in-depth latency reports
• Can collect data from App Engine, Google HTTP(S) load balances and applications implementing Cloud Trace SDKs

Stackdriver Debugger

• Let’s us inspect the state of a running application in real-time without stopping it or slowing it down significantly
• It can capture debug snapshots, call stack and local variables of a running application