Monitoring Support for Manta Flow Agent in Cloud-Based Architecture
| Thesis title in Czech: | Podpora monitorování Manta Flow agentů v cloudové architektuře |
|---|---|
| Thesis title in English: | Monitoring Support for Manta Flow Agent in Cloud-Based Architecture |
| Key words: | monitoring|time series databázy|Java inštrumentačný agent|multi-agentné prostredie|cloud architektura |
| English key words: | monitoring|Java instrumentation agent|multi-agent environment|cloud architecture|time series databases |
| Academic year of topic announcement: | 2021/2022 |
| Thesis type: | diploma thesis |
| Thesis language: | angličtina |
| Department: | Department of Distributed and Dependable Systems (32-KDSS) |
| Supervisor: | doc. RNDr. Pavel Parízek, Ph.D. |
| Author: | Mgr. Roman Firment - assigned and confirmed by the Study Dept. |
| Date of registration: | 29.06.2021 |
| Date of assignment: | 30.06.2021 |
| Confirmed by Study dept. on: | 09.07.2021 |
| Date and time of defence: | 07.06.2022 10:00 |
| Date of electronic submission: | 05.05.2022 |
| Date of submission of printed version: | 16.05.2022 |
| Date of proceeded defence: | 07.06.2022 |
| Opponents: | Mgr. Filip Kliber |
| Guidelines |
| The next major version of the Manta Flow platform for data lineage analysis will support cloud-based architecture where only the necessary components, in particular the Manta Flow Agent with extractor plugins, should run in the customer's environment (on-premises), while all other components (e.g., the data lineage analysis itself) will be deployed somewhere in the cloud.
In this context, a very needed feature is the support for monitoring of the Manta Flow Agent, its status, communication with other components of the Manta Flow platform, and interaction with internal systems of the respective customer. The goals of this master thesis project are the following: (1) analyze functional requirements on the monitoring framework for Manta Flow Agent, (2) analyze possible use cases of running multiple agents in the environment of a single customer, focusing on the management of agents and their interaction, (3) then propose a solution that fits into the Manta Flow platform and reflects also important non-functional requirements (performance overhead, security risks), and (4) finally create a proof-of-concept implementation just for the scenario of a single agent for one customer. Results of the analysis will include the list of relevant metrics and discussion of possible ways how to retrieve their values. The monitoring framework should also observe how efficiently agents (and their plugins) use resources in a computer system, that means CPU, memory usage, disk space, and network bandwidth, for example. Author should use existing robust technologies and libraries for runtime monitoring of Java programs, such as JVM TI and JMX, and for bytecode instrumentation. |
| References |
| 1. Java Virtual Machine Tool Interface (JVM TI), https://docs.oracle.com/javase/8/docs/platform/jvmti/jvmti.html
2. Java Management Extensions (JMX), https://docs.oracle.com/javase/8/docs/technotes/guides/jmx 3. Java agent and package java.lang.instrument, https://docs.oracle.com/javase/8/docs/technotes/guides/instrumentation/index.html 4. Walter Binder, Jarle Hulaas, and Philippe Moret. Reengineering Standard Java Runtime Systems through Dynamic Bytecode Instrumentation. SCAM 2007 5. Luis Mastrangelo and Matthias Hauswirth. JNIF: Java Native Instrumentation Framework. PPPJ 2014 6. Holger Eichelberger and Klaus Schmid. Flexible resource monitoring of Java programs. Journal of Systems and Software, volume 93, 2014 |