Feature: Custom metrics must up to date when change occurs on related data

[hocinehacherouf]

Is your feature request related to a problem? Please describe. Currently, custom metrics are calculated each time interval (by default each 10 minutes). Between each iteration, these metrics can be outdated.

Describe the solution you'd like To keep custom metrics up to date: When a data has been created, updated, deleted, its related custom metric must be recalculated unitary

[kbeaugrand]

Actions :

• [ ] MetricLoaderService
• [ ] Add Each method by device kind metrics (devices, concentrators, edgeDevices)
• [ ] Call correct method in CRD methods from API controllers

[hocinehacherouf]

@kbeaugrand currently we use scheduled jobs to retrieve metrics about devices from azure iot hub. Some metrics can be queried from our database, other metrics can be retrieved only on azure iot hub.

I propose:

• Add/sync missing data on our database (example: [data] connection state of edge devices ==> [metric] count of connected edge devices) to avoid any dependency with azure iot hub
• These metrics can be recalculated by each call to the metric endpoint (with cache) OR reusing the existing metrics job loader with short sync interval.

[kbeaugrand]

Saw in PBR.

No problem to synchronize the data during device sync job. Metrics might be taken directly from our database. Add the cache at the controller level.

This will make the settings around metrics job useless (important change not breaking).