Skip to content

deib-polimi/modaclouds-monitoring-manager

Repository files navigation

The MODAClouds Monitoring Manager

The monitoring manager was migrated and included in the new Tower 4Clouds Repository


In the context of MODAClouds European project (www.modaclouds.eu), Politecnico was one of the partners involved in the development of the Monitoring Platform.

The monitoring manager is the main coordinator of the platform. It is responsible of installing monitoring rules, configuring monitoring components, attaching external observers to requested metrics and keeping the knowledge base up to date through the interaction with Models@Runtime. The monitoring manager is also is the main interface towards external components.

Please refer to deliverable D6.3.2 to better understand the role of this component in the MODAClouds Monitoring Platform.

Refer to the Monitoring Platform Wiki for installation and usage of the whole platform.

Documentation

Take a look at the documentation table of contents.

Change List

v1.7:

v1.6:

  • added a Vagrantfile in the root of the repo for testing purpose, see the developer manual
  • added a Vagrantfile in the released assembly for launching the entire monitoring platform using Vagrant, see the user manual
  • fixed problem that did not prevent attaching observers with malformed URL
  • added initial structure for the webapp
  • fixed problem which did not exclude OutputMetric actions from new actions
  • improved logging and error messages
  • improved validation when updating/uploading the deployment model
  • fixed problem that caused errors when uninstalling rules that were using the same input metric
  • implemented test observers and included in the released assembly, see the developer manual and user manual
  • updated to knowledge-base-api 2.3.1
  • updatad to data-collector-factory 0.3.3
  • updated to qos-models 2.4.1

v1.5:

  • integrated with qos-models 2.4 for actions execution
  • a private endpoint was implemented for internal communication and the port can be specified through the configuration, see the user manual
  • implemented a new REST API for retrieving all resources in the model (GET /model/resources), see the doc

v1.4:

  • Aggregations and mathematical expressions can now be expressed in the OutputAction value parameter.
  • The monitoring manager waits for the availability of the DDA and KB for 30 seconds before failing the startup
  • Both DDA and KB are now cleaned and reset automatically at monitoring manager startup
  • All REST api were bug fixed and are now working
  • REST api changes: both GET /metrics and GET /metrics/{id}/observers now return a json file, see the doc
  • KB and DDA are now run both on a single vagrant machine during integration test
  • updated to csparqool 1.2.2
  • updated to knowledge-base-api 2.2.1
  • updatad to data-collector-factory 0.3.2
  • updated to qos-models 2.3

v1.3.4:

  • the default list of metrics for monitoring rules can now be overwritten by a custom one, see the doc

v1.3.3:

v1.3.1:

v1.3:

  • implemented API: GET /model/resources/:id
  • updated to csparqool 1.2.1
  • updated to knowledge-base-api 2.2
  • updatad to data-collector-factory 0.3
  • updated to qos-models 2.2
  • queries are now created according to the new qos-models 2.2 version of monitoring rules:
    • metricAggregation is optional
  • outputMetric action in monitoring rules now accepts 3 parameters (resourceId, metric, value), check the updated list
  • sda are now just observers & data collectors from the monitoring manager perspective
  • bug fixes

v1.2:

  • internalComponents can now be a monitoredTarget
  • system properties can be used now besides environemnt variables
  • inheritance was removed

v1.1.2:

  • Fixed a bug that didn't allow to upload a new model the first time using the PUT method
  • updated to qos-models 2.1.4: check qos-models change list for the new of version of monitoring rules
  • installation instructions updated (look at the documentation)