New Project: Hubitat System Metrics (Node Red, SQL Server, Grafana)


#1

New Project: Hubitat System Metrics (Node Red, SQL, Grafana)

Objective: Create a storage and reporting system for all configured exportable data, events and logs.

Calculate historical system metrics across all system component domains.

  • Application Level Logs
  • Device Logs
  • System Logs (all data)
  • Event Logs (all data)
  • All logs are captured and pinned to local time, with environmental #s (temp, humidity, etc..) as well as configurable Secondary and Tertiary Input devices. You pick the two devices, the commands and configure the responses headers.
  • Archive system backups as BLOB into the database (extreme waste of storage space? Let's see)

Reporting: Grafana and possibly tabular output using a web based Crystal Report configured dynamically system devices, applications,etc...

SQL Server: To be honest I have to do a little leg work to determine the storage capacity requirements for a "larger" system, exporting 100% of all available log activity stored 24/7/365

Side-note: The completion of this project assumes my Hubitat doesn't have a stroke with all the extra logging turned on.

For some time I experienced what I interpreted as a "bogged down system"
After digging in I realized I had a LOT of logging features turned on, as well as many devices needing clean-up from dropped devices to non-functional z-wave repeaters , broken rules ,etc..

Either way.. Those problems have all been long resolved. What caused the bogged down system initially? IDK. I didn't address each problem and retest the system performance. I will fully evaluate if that problem were to return.

Matt


#2

Sounds good. Have you read this post [Hubitat Data Collection] ?
(Hubitat > NodeRed > MySQL > Grafana (LONG READ))


#3

Yes. Many times


#4

There is almost zero extra load on the hub, as events and logs are already exposed on websockets.

You might consider influxdb over sql, as its a time-based data store with auto-purging of old data, so you dont have to worry about filling up your disk. grafana integrates with influxdb so easily.


#5

Would have though so but was just to be sure :slight_smile:


#6

I have been using SQL for a long time. It is easily pulling in that much data live