Integrating Time Series-based Monitoring with Run-time Modelling: Unterschied zwischen den Versionen

Aus SDQ-Institutsseminar
(Die Seite wurde neu angelegt: „{{Vortrag |vortragender=Eduard Kukuy |email=utehb@student.kit.edu |vortragstyp=Bachelorarbeit |betreuer=Robert Heinrich |termin=Institutsseminar/2019-10-25 |ku…“)
 
Keine Bearbeitungszusammenfassung
Zeile 5: Zeile 5:
|betreuer=Robert Heinrich
|betreuer=Robert Heinrich
|termin=Institutsseminar/2019-10-25
|termin=Institutsseminar/2019-10-25
|kurzfassung=TBD
|kurzfassung=Cloud systems may consist of collections of smaller software components (in some cases called microservices), possibly written in different programming languages and hosted across various hardware nodes. These components require continuous adaptation to changing workload and privacy constraints. There exist approaches solving this problem already, but they come along with limitations including binding to a certain platform or programming languages and not accurate handling of multi-host applications.
 
This thesis presents an approach to platform-independent observing of cloud applications, including comprehensive monitoring of relationships between components of the system. The concept of a time series database is used under the hood for storing monitoring data. It gets then transformed into the format needed for the performance model extraction.
Furthermore, a complete specific implementation of the approach with exemplary tools is provided.
}}
}}

Version vom 4. Oktober 2019, 08:57 Uhr

Vortragende(r) Eduard Kukuy
Vortragstyp Bachelorarbeit
Betreuer(in) Robert Heinrich
Termin Fr 25. Oktober 2019
Vortragsmodus
Kurzfassung Cloud systems may consist of collections of smaller software components (in some cases called microservices), possibly written in different programming languages and hosted across various hardware nodes. These components require continuous adaptation to changing workload and privacy constraints. There exist approaches solving this problem already, but they come along with limitations including binding to a certain platform or programming languages and not accurate handling of multi-host applications.

This thesis presents an approach to platform-independent observing of cloud applications, including comprehensive monitoring of relationships between components of the system. The concept of a time series database is used under the hood for storing monitoring data. It gets then transformed into the format needed for the performance model extraction. Furthermore, a complete specific implementation of the approach with exemplary tools is provided.