Hubitat Easy Dashboards with Graph tiles
These are pretty easy to setup. You simply create a dashboard using Easy Dashboards and add the appropriate tile. Once the tile is created you select the tile type to be graph and what measurement you want to include. This will display a graph for that metric.
Pro’s
- Extremely easy to implement since it uses the Easy Dashboards and built in tiles
- Built into the hub firmware and is continuing to get minor improvements.
- Runs completely on Hub
Con’s
- Very simplistic with no advanced functions.
- Only displays a line graph or bars.
- Only uses metrics stored on hub which tend to be limited
Use Case
This is a reasonable solution if you just want something to show current metrics. This will likely create a basic picture of how the hub is running in a recent interval but will not provide general trends of what is going on over significant time. The limitation of only showing based on stored event history an be very limiting. That said this can be a decent place to start.
Webcore with Hubigraphs add-on
I am aware of the solution and it has a fallowing. If you use Hubigraphs before this is a good option for you.
Pro’s
- Completely local to hub graphing solution
- Graphs work with built in dashboards
- If you already have Webcore, you are already part of the way there
- Good Community support.
Con’s
- Long term storage on hub in hub storage.
- Takes up hub resources when graphing or event tracking
- This was effectively a port of Hubigraphs that was not getting active updates.
- Takes a bit of fiddling to figure out how to get them to display what you want
Use Case
Honestly i need help with this. This is not a solution i have used. Without hands on experience, i would expect if you already use Webcore this could be a good option.
ConstantGraph
ConstantGraph appears to be a free/paid solution that is cloud based that a user would sign up with for a curated solution to display various metrics from the Hub. The free tier seems to be very similar to Watchtower except for the use of event data vs 5 min intervals for the lowest precision of records.
Pro's
Free and paid solutions which means very easy maintance.
Currated dashboards which are nice, predefined, and functional
Cons
Cloud Dependent.
Potentially additonal cost if you want more data then what the free plan provides
Use Case
This could be good for the user that wants to get the graphing function off the Hub, and is satisfied with something similar to Watchtower.
Watchtower
Watchtower is a fully local graphing solution that canl provide a good overview of the device statistics. Though it runs on the hub Watchtower implements downsampling. This means it takes steps to limit the number of records and size of the stored data by keeping representational data for given intervals.
Pro’s
- Completely local to the hub
- Performs downsampling to reduce storage and cpu used when accessing older data
- Has Long term storage that can be used to show trends over a long period of time.
- Creates its own dashboards for the metrics with some very interesting configuration options.
- Actively being developed by the Community developer
Con’s
- Because it uses intervals to poll data and does downsampling the accuracy of the data should be considered rounded for the interval the data is being displayed for.
- Keeps data locally on hub which means it is using local IO, CPU, and Memory to operate.
- Maximum accuracy is 5 min polling. Will not provide more accurate information.
- Because downsampling is used this is primarily for displaying numerical metrics. May not be optimal for State metrics, but does have the ability to support them.
Use Case
This is a very good flexible solution for those that want an all local solution that you can set and forget. The automated downsampling in many ways is a fantastic option many other solutions are missing and really helps keep the size of the retained data down.The lack of accuracy is only an issue when you find a use case that needs it and as such this solution will meet most users needs easily
Grafana, Influxdb, Stack
The use of Grafana and Influxdb is by far the most powerful, robust, and flexible option. At the same time it is easily the most complex. It requires you to either upload your data to the cloud, or run an always on host in your home to collect the events from the hub as they occur. A smart app called InfluxDB Logger is uses in most cases to upload event data to the Influxdb database. The latest version of Influxdb v3 allows MQTT to be used to upload events into the DB. Once the events are in the database, Grafana is an enterprise level graphing, monitoring, and analytics tool used to take the raw metrics and create visual representation of your hubs events.
Pro’s
- Influxdb is a purpose built Time Series database, and as such is extremely performant
- Most robust graphing and alert solution with Grafana with it’s various panel options
- Completely removes hub from picture so more advanced graphing functions can be performed
- Because you are running this on external hardware, storage and IO concerns are no longer an issue
- A good variety of prebuilt dashboards exist already
- Because you can set this up on fast storage can keep full accuracy data for much longer then if the data resided on the hub.
- Grafana is smart enough to request data in appropriate intervals to only retrieve data it needs so downsampling isn’t necessary.
Con’s
- Requires always on computer, or uploading your data to the cloud.
- Because this is a robust solution your setup will like require decent amount of resources
- You still need to be aware of the resources needed to run on those external systems
- Downsampling data is not automatic and you will need to manage how it is done.
- To manage storage usage you may need to setup retention limitations
- IT investment for managing always on gear if you run it locally at home
- Dashboards require you to know how to query the database which may not be straight forward. Starting dashboards from scratch may not be easy.
Use Case
Simply put this is a great option if you already run a home lab or want the most accurate database possible while not being worried about a little upkeep. In general once setup this solution will just run with little concern. Then when you want the most complete data possible Grafana provides an extremely dynamic tool to display any metrics you want. Having this off the hub you can res assured it won’t impact anything running on the hub, and grafana can be made available anywhere.
I wanted to start a thread to discuss the various options for Graphing with Hubitat. I can personally think of 4 that i have either tested somewhat or read about so i started with the above post of information. If you can think of more options, pro’s, con's for each one let let me know and I will update the list above. The idea here is to simply inform folks. If i am wrong or need to elaborate on something above please let me know so i can update it.
