How do you ship code faster and reliably without making sacrifices? Ben Mostafa shares actionable insights into driving DevOps practices in the real world.
14. Change lead time
What to measure?
1 Time spent in each step
2 Wait time between each step
3 Pull request metrics – time to close, reviewers/approvals, comments and
size
15. Change lead time
What to measure?
1 Time spent in each step
2 Wait time between each step
3
Pull request metrics – time to close, reviewers/approvals, comments and
size
16. Change lead time
What to measure?
1 Time spent in each step
2 Wait time between each step
3
Pull request metrics – time to close, reviewers/approvals, comments and
size
18. Wait Time vs %
resource busy
Wait Time = (% Busy)/(% Idle) Wait Time
19. Codebase health
What to measure?
1 Age of dependencies/libraries
2 Age of vulnerabilities/libraries with vulnerabilities
3 Complexity, duplications, test coverage etc…
20. Codebase health
What to measure?
1 Age of dependencies/libraries
2 Age of vulnerabilities/libraries with vulnerabilities
3 Complexity, duplications, test coverage etc…
21. Codebase health
What to measure?
1 Age of dependencies/libraries
2 Age of vulnerabilities/libraries with vulnerabilities
3 Complexity, duplications, test coverage etc…
22. Codebase
health
Dependent Library Age
Jackson
JSON
AWS SDK
Apache Commons
IO
Log4J
B
Overall
codebase health
3m
Average
dependency age
10
Known
vulnerabilities
Repository 1
Repository 2
Repository 3
Repository 4
Repository 5
Repository 6
B
A
D
F
A
A
23. Service health
What to measure?
1 Data stores and communication channels
2 Deployability
3 Alignment with architectural principals
4 Time since last deployment
24. Service health
What to measure?
1 Data stores and communication channels
2 Deployability
3 Alignment with architectural principals
4 Time since last deployment
25. Service health
What to measure?
1 Data stores and communication channels
2 Deployability
3 Alignment with architectural principals
4 Time since last deployment
26. Service health
What to measure?
1 Data stores and communication channels
2 Deployability
3 Alignment with architectural principals
4 Time since last deployment
27. Service Health
score card
Attribute Weight Score
TOTAL
Is HTTPS used for all
HTTP communication?
Is data encrypted at rest
Does this service handle more
than one concern?
Is this service stateless?
Can this service be deployed
without downtime / service
interruption?
Can this service be scaled
horizontally?
Is the data decoupled
from the applications?
How long since the service was
last deployed to production?
10 10
10 0
20 20
10 10
20 0
10 10
5 5
-0.1 -2
53
C
Overall Service
Health
Score
75 - 85
60 - 75
50 - 60
35 - 50
Rating
A
B
C
D
28. What else
to measure?
• Age and vulnerabilities in 3rd party software
• Platform utilisation
• Operational alerts per team or service
• Team health
29. Recap
Measure
deployments
• Deployments per
engineer
• Deployments duration
• Time of day
• Success/failure
Measure
change lead time
• Time spent in each step
• Wait time
• Pull request metrics
Codebase
health
• Age of dependencies
• Age of vulnerabilities
• Overall health rating
Service
health
• Data stores and
communication channels
• Deployability
• Alignment with architectural
principals
• Time since last deployment