Engine monitoring inperformant

[simonhir]

Describe the bug

The modification and improvement of the engine monitoring within #1229 is currently using very much memory when running. The problem seems to be the large amount of process definitions.

To Reproduce

• See current avg mem usage of engine
• Enable i.e. task monitoring with property DIGIWF_PROMETHEUS_PROCESS-ENGINE_PROVIDERS_TASK: 'true'
• Engine uses about 0.5 Gb more memory

Expected behavior
The additional needed resources when enabling Prometheus monitoring is as low as possible and doesn't restrict the engine performance while delivering all necessary information for a good monitoring.

Acceptance criteria

• metrics can run without using too much resources or having impact on engine performance

[simonhir]

Ops-Mr: https://git.muenchen.de/digitalisierung/digiwf-ops/-/merge_requests/417