Monitoring the right
way
Dan Barker
@barkerd427
● Metrics aggregation
● Log aggregation
● Alerting and visualizations
● Distributed Tracing
Observability Tools
@barkerd427
● Metrics aggregation
● Log aggregation
● Alerting and visualizations
● Distributed Tracing
Observability Tools
@barkerd427
● Easy to get started
● Install an agent
● Built in
Metrics Aggregation
@barkerd427
● Some built in
● Requires dev work
● Security risks
Log Aggregation
@barkerd427
● Requires more
understanding
● Can cause negative
impact
Alerts and Visualizations
@barkerd427
● Not necessary
● Not built in
● Less mature
Distributed Tracing
@barkerd427
Push vs Pull?
@barkerd427
OpenMetrics
@barkerd427
Use Business
Metrics
@barkerd427
Log aggregation
tools are for logs!
@barkerd427
● actually have logs
● format in JSON
Do’s of logging
@barkerd427
● include a timestamp
● time should be UTC
Do’s of logging
@barkerd427
● log all application errors
● write messages in a
human-readable form
Do’s of logging
@barkerd427
● log info data in prod
● log insignificant events
● log PII
Don’ts of logging
@barkerd427
“A joke that works is
complete knowledge in a
nanosecond.”
– Steve Martin
@barkerd427
There are no
“informational”
alerts!
@barkerd427
OpenTracing
@barkerd427
OpenCensus
@barkerd427
● https://monitoring.danbarker.codes
● https://www.practicalmonitoring.com/
● https://opensource.com
● https://danbarker.codes
Resources
@barkerd427

Monitoring the right way - DevOpsDays Kiev - Ignite