openstack
/
monasca-analytics
Code Issues Proposed changes

Update README 

Currently, README has invalid markup which will not be rendered on PyPI.
This patch updates README to be rendered on PyPI and to be more readable.

Change-Id: Ic081c0729e44f4e09ac21c47a5dbbd0da6deb77f
This commit is contained in:
Daisuke Fujita 2019-01-25 04:11:07 +00:00
parent 1d9583af98
commit 474e4b8917
2 changed files with 23 additions and 22 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
README.md
View File

@ -23,37 +23,18 @@ we can speed up the extraction of actionable infrastructure insight.
### Advantages
:thumbsup: Decouple algorithm design from execution.
:thumbsup: Reusable specification of the desired
[flow](doc/design.md#flow).
:thumbsup: Reusable specification of the desired [flow](doc/design.md#flow).
:thumbsup: Language independent [flow](doc/design.md#flow) definition.
:thumbsup: Data source and format independent.
:thumbsup: Easy to add new [SML](doc/design.md#sml) algorithms and #
combine them with pre-existing ones in the [flow](doc/design.md#flow).
:thumbsup: Easy to add new [SML](doc/design.md#sml) algorithms and # combine them with pre-existing ones in the [flow](doc/design.md#flow).
:thumbsup: Transparently exploit data parallelism.
### Example Use Cases
Below are few use cases that are relevant to OpenStack. However, MoNanas
enables you to add your own [data ingestors](doc/dev_guide.md#ingestors).
| Example | Alert Fatigue Management | Anomaly Detection |
|:------------------------------|:-------------------------|:------------------|
| **Dataset** | Synthetic, but representative, set of Monasca alerts that are processed in a stream manner. This alert set represents alerts that are seen in a data center consisting of several racks, enclosures and nodes. | `iptables` rules together with the number of times they are fired in a time period. |
| **Parsing** | Monasca alert parser. | Simple parser extracting period and number of fire events per rule. |
| **SML algorithm flow** | `filter(bad_formatted) -> filter(duplicates) -> aggregate() >> aggregator` aggregation can utilize conditional independence causality, score-based causality, linear algebra causality. | `detect_anomaly() >> aggregator` anomaly detection could be based on SVM, trend, etc. |
| **Output** | Directed acyclic alert graph with potential root causes at the top. | Rule set with an anomalous number of firing times in a time period. |
| **:information_source: Note** | Even though this could be consumed directly by devops, the usage of [Vitrage MoNanas Sink](doc/getting_started.md#vitrage_sink) is recommended. The output of this module can speed up creation of a [Vitrage](https://wiki.openstack.org/wiki/Vitrage) entity graph to do further analysis on it. | None. |
`->` indicates a sequential operation in the flow.
`//` indicates beginning of group of operations running in parallel.
`-` indicates operations running in parallel.
`>>` indicates end of group of operations running in parallel.
* [MoNanas/UseCases](doc/use_cases.md): Use cases that are relevant to OpenStack
### Documentation
* [MoNanas/GettingStarted](doc/getting_started.md): A starting point for users

20
doc/use_cases.md Normal file
View File

@ -0,0 +1,20 @@
# Example Use Cases
Below are few use cases that are relevant to OpenStack. However, MoNanas
enables you to add your own [data ingestors](doc/dev_guide.md#ingestors).
| Example | Alert Fatigue Management | Anomaly Detection |
|:------------------------------|:-------------------------|:------------------|
| **Dataset** | Synthetic, but representative, set of Monasca alerts that are processed in a stream manner. This alert set represents alerts that are seen in a data center consisting of several racks, enclosures and nodes. | `iptables` rules together with the number of times they are fired in a time period. |
| **Parsing** | Monasca alert parser. | Simple parser extracting period and number of fire events per rule. |
| **SML algorithm flow** | `filter(bad_formatted) -> filter(duplicates) -> aggregate() >> aggregator` aggregation can utilize conditional independence causality, score-based causality, linear algebra causality. | `detect_anomaly() >> aggregator` anomaly detection could be based on SVM, trend, etc. |
| **Output** | Directed acyclic alert graph with potential root causes at the top. | Rule set with an anomalous number of firing times in a time period. |
| **:information_source: Note** | Even though this could be consumed directly by devops, the usage of [Vitrage MoNanas Sink](doc/getting_started.md#vitrage_sink) is recommended. The output of this module can speed up creation of a [Vitrage](https://wiki.openstack.org/wiki/Vitrage) entity graph to do further analysis on it. | None. |
`->` indicates a sequential operation in the flow.
`//` indicates beginning of group of operations running in parallel.
`-` indicates operations running in parallel.
`>>` indicates end of group of operations running in parallel.