Monitor and control all system processes and environments
|Actors||Integrated Observatory Operator (Cyber Operator)|
|References||UC.R2.40 Monitor ION Resources|
|Uses||UC.R2.44 Define Service Type During Runtime|
|Is Used By||UC.R2.46 Operate Integrated Observatory Network|
|Is Extended By|
|In Acceptance Scenarios||AS.R2.01C Operate Integrated Observatory Network|
|Primary Service||Process Management Services|
This information summarizes the Use Case functionality.
Monitor all processes in the ION system (services, agents, user processes) and their deployment execution environment. Control process scheduling policy, executing processes, and execution sites.
Integrated Observatory Operator is logged in.
- The Integrated Observatory Operator calls up a Process Collection view to review process environments on the system.
- See UC.R2.49 Deploy Distributed Processes for further details.
- Additional detail: The displays should be able to discriminate/filter based on process type (a service, an agent), owner class (system, user), and specific owner.
- <2> The Integrated Observatory Operator calls up detailed Process Resources page views to examine resource usage for a class of processes.
- The ability to drill down from the ION level, through site, availability zone, Execution Engine and examine resource usage at each level is important.
- The ability to filter on different types of resources (CPU, CCs, connections, message I/O, message queue length) is important.
- If this information is not available through the Integrated Observatory in R2, it can be viewed from outside the system.
- Upon identifying a runaway process (using too many resources of some type), the Integrated Observatory Operator can take appropriate action.
- Could pause data input, or kill it.
- Theoretically could put it into debug mode, or move it to an isolated machine, but moving a process to an isolated machine with current in-flight context is hard.
- The Integrated Observatory Operator examines the scheduling policy for a process, and modifies it.
- Could simply increase or decrease the frequency of execution, for those processes that are scheduled by time interval.
- Could increase or decrease the priority allocated to the process (e.g., unix process priority in the context of the operational unit).
- Could create a more advanced policy, that defines the scheduling based on other parameters (CPU time used by process or machine, for example).
- <2> The Integrated Observatory Operator examines the scheduling policy for a group of processes, and modifies them all.
Operations continue with the changes in effect.
These comments provide additional context (usually quite technical) for editors of the use case.
Integration of the CEI processes with the Integrated Observatory COI resources infrastructure is under analysis. The ability to manage these processes from the Integrated Observatory is therefore not yet determined.
The level of process inspection within the Integrated Observatory that's required in Release 2 may need further analysis. The scenario assumes that detailed process inspection will take place outside of the Integrated Observatory.
There is some duplication of steps with UC.R2.40 Monitor ION Resources, but the details under each step are somewhat different.