1 of 28

Reliability Dashboards

January, 2019

Presentation Name

Month 00, 2018

1

2 of 28

Anomaly Detection & Root Cause for QA, DevOps and SREs.

OverOps Reliability Dashboards

An open-source project that enables integration of OverOps data via REST APIs.
Machine Learning detect anomalies for apps, deployments and infrastructure tiers across environments.
A set of Grafana dashboards provide visualization and advanced drill down capabilities.
A Jenkins plugin provides reliability report and active quality gates in CI/CD.
All provided widgets can be integrated into any existing operational dashboard for data correlation.

2

3 of 28

The Reliability Scorecard dashboard provides an overview of all anomalies detected within a target deployment, application or infrastructure tier within the selected environment(s), assigning each its own dynamic score.

Anomalies include:

* Newly introduced errors.

* Regressed / increasing errors.

* Performance slowdowns.

Each anomaly is assigned a Severity, and can be drilled into to see its Root cause.

Learn more about scoring here.

3

4 of 28

4

5 of 28

The Deployments pane shows the state + score of each active deployment within the selected environment(s).

5

6 of 28

Hovering over New errors will show their code locations. Severe errors are highlighted. Click an error to drill into its Root Cause.

6

7 of 28

The New errors drill-down shows all new errors new to a deployment, application or infrastructure tier. Click an error to see its Root cause.

7

8 of 28

Hovering over Increasing Errors will show their origin and rate of increase(%). Severe regressions are highlighted. Clicking an error drills into its Root cause.

8

9 of 28

Hovering over new errors will show their locations. Severe (P1) errors are highlighted. Clicking an error will jump to its root cause analysis.

The Increasing Errors drill-down shows the increase in volume and percentage of the regression, comparing it to a dynamic baseline. Clicking an error will show its root cause.

9

10 of 28

The Root Cause drilldown shows the state of the code at the moment of regression across the entire call stack - 10 levels into the heap.

Time of Regression

10

11 of 28

Hovering over Slowdowns shows their rate change. Severe slowdowns are highlighted. Clicking a slowdown drills into its Root cause.

11

12 of 28

The Slowdowns drill-down shows the increase in response time for each slowing transaction compared to a dynamic baseline. Click to see the Root cause of the slowdown.

12

13 of 28

Auto Timers show the state of the code at the moment of slowdown across the entire call stack - 10 levels into the heap.

Location of slowdown highlighted.

13

14 of 28

Also captured are the full environment state of the container or instance, and last 250 lines of DEBUG-level log statements.

14

15 of 28

The Tiers pane shows the score of each Infrastructure Tier used within the environment(s). Tiers are automatically identified via dynamic code analysis.

15

16 of 28

The Applications pane shows the score of each microservice or monolith. Click each one to drill down to Reliability Analysis.

16

17 of 28

Reliability Analysis shows Anomalies in a target app, deployment or infrastructure tier. Anomalies are prioritized based on volume, regression and origin in code. Clicking a regression or slowdown shows its Root cause.

17

18 of 28

18

19 of 28

The Performance dashboard shows the state + score of every transaction within the system. Clicking a transaction will show the Root cause of its slowdown.

19

20 of 28

20

21 of 28

Each Transaction shows the error occuring within it, including the volume and rate each error. Click to drill-down.

21

22 of 28

The Errors drill-down shows errors within target transaction(s). If a Jira issue has been opened, it can be accessed directly. Click an error for its Root Cause.

22

23 of 28

The Root Cause drill-down shows the code and variable state at the moment of error across the entire call stack - 10 levels into the heap, as well environment state and DEBUG-level statements.

23

24 of 28

The Event Diff dashboard enables QA and DevOps/SRE teams to compare between different releases.

24

25 of 28

As well as different time periods.

25

26 of 28

Promotion Gates in Jenkins apply Machine Learning to extend code coverage by automatically marking a build as unstable.

OverOps provides four code Quality Gates for CI/CD environments, based on:

Error Volume
Unique Error Count
Critical New Issues
Critical Regressions

26

27 of 28

The Jenkins plugin integrates into your CI/CD environment to report on new issues and regressions during the test and integration phase.

27

28 of 28

Thank you

Presentation Name

Month 00, 2018

28