Stashfin Metrics
Management Metrics / Tools
Gitlab / Code
- Number of repositories
- Size of repositories
- Number of commits and MRs
- Test cases
- CodeCoverage
- SonarQube - Smells and Errors / CodeClimate
- Number of bugs in module
- Sentry alerts
- ECR Images
- Gitlab
- Response times of modules
- Jenkins
- Number of CICD Pipelines
- Number of deployments
- Contribution Analytics (Bronze paid version gitlab) https://docs.gitlab.com/ee/user/group/contribution_analytics/
Grafana
- Istio Service Mesh
- Response Time
- Tail latencies (95th percentile)
- newrelic dashboard
- Pod usage
- DB Audit logs
- CloudFlare metrics
- Loki - logs - search for error
Databases
- AWS / MySQL WorkBench Performance
- Top most high cost sql queries
- Table Size
- Index sizes
- Number of tables
AWS
- Billing review
- IAM Users / configurations
- Resources
Key Metrics
Concurrency
Current - 3500 req/min = 60 req/sec Scalable - 500 req/sec 10 cr users - 10 cr / 3000 - 33333 users With lead to approval ratio of 10%, you need - 3,33,333 Tech scalable for 8 lakh users
Write key metrics of stashfin (peak workloads till now and expected workloads)
- number of customers (active customers)
- number of loans processed daily / monthly
- % of people who default goes to collection process
- number of daily email sent
- number of marketing email sent
- number of daily / monthly sms sent
Business Metrics / PowerBI
PowerBI Dashboard / Reports
- Communications
- Payment Inflow Dashboard
- Funnel SF
- Elevate Daily Tracker
- Daily Summary
- Funnel Reports
- TAT Reports
- Elevate Daily Tracker
- General Funnel - Monthly disbursal
Other Metrics
- Vendor Billing external API
- Capacity planning
- Add all the metrics value
- Revised ETA column in sheet
Core Metrics
1. Devops
- Uptime to be 100% (less than 45 mins a month of TSO (Total System Outage) & less 2 hours of MSO & Less than 4 hours of SSO)
- Infra cost on a unit loan disbursed to be reduced by 20% QOQ - Pls publish current (CPU, Storage, Overall)
- Website / API Response time (Website less than 3 seconds on all platforms, API's less than 500ms)
- 4XX / 5XX to come down by 20% MOM
- Complete metering of all External API calls to tally 100% with Vendor billing
- 90% ETA compliance on tasks
- 100% compliance of Post incident analysis on all issues within 24hrs including actions items to be put in Dev Q
- Publish a scale out plan & Infra budget based on sales growing MOM 20% & adhere to it in actuals
2. Engineering
- DB performance to improve by 15% MOM (net CPU to go down / net cost to go down & choking issues to reduce)
- Sonar issues to come down by 20% MOM
- Code Complexity to reduce by 20%
- 50% of outages reported should have a preemptive time stamp on engineering side (we should know before users)
- 100% test case coverage of user journey for all customers
- All user facing properties to be tested 10+ times each day
- Optimizations
- 2 New implementations each months for Scalability enhancements
- 2 New implementations each months for Security
- 2 New implementations each months for Confidence / Testing / Alerting
3. Tech PMO
- Number of Projects per week (New, In Progress, Complete)
- Cut by Adhoc, Project & Bugs
- ETA Compliance to be >95% (Net of Dependencies)
- All dependencies to be communicated within first 20% of Project cycle
- All projects to have Infosec sign off
4. Tech QA
- Measure weekly production bug count
- Test case coverage to be >95%
- Test case automation off coverage to be >75%
- 100% API documentation on PostMan & Runners
5. Decision Engine
- 100% Monitoring of DE Daily
- All cases (100%) to be decisioned with 0% miss & 1% (or below) error rate
- 100% documentation of all changes in DE
- 10 cases in Libr8 FPD & 20 cases in Elev8 Non starter to be reviewed & feedback posted each week