reef-pi :: An opensource reef tank controller based on Raspberry Pi.

[Ranjib]

Regarding mqtt, do we have a GitHub issue to track it ? If not we should start with that and slowly get it fixed. I don’t use mqtt hence not on top of those issues

 

[Ranjib]

5.1. I've deleted ato and added back an ato. Let it run for a while and get graph data. Still same error when trying to reset ato.

Can't reset any ato.

This was minor UI bug. I have fixed it already and will be out in 5.2 . Super sorry for overlooking this and thank you so much for reporting it.

 

[elysics]

Funny you say that , originally it was not there , but that created confusion during the December to January transition and I had to update code to include year.
this is very simple visual change that we can always change, it’s more about the user experience we have to settle in. I’m open for suggestions

I thought about this before, would it be possible to make what is displayed dependant on the time-width of the graph? I.e. only display hours and minutes OR day and month OR just month OR month and year OR just year depending on how much time the graph covers. The full timestamp can still be left in the popup whn selecting datapoints to clear up confusions

Or something similar, to declutter irrelevant data

 

[marekd1]

Funny you say that , originally it was not there , but that created confusion during the December to January transition and I had to update code to include year.
this is very simple visual change that we can always change, it’s more about the user experience we have to settle in. I’m open for suggestions

LOL... I understand the reasoning to include it but I if you have month as you do I am not clear on confusion. I suspect if you want year history graph I would expect that folks are using options like Grafana. In any event are there any possibilities to include an option/switch include year or do not include year this might be the easiest way to make everyone happy ? :)

or better yet love the elysics suggestion above

 

[marekd1]

Regarding mqtt, do we have a GitHub issue to track it ? If not we should start with that and slowly get it fixed. I don’t use mqtt hence not on top of those issues

Ok I will post an issue to track it. I use Grafana extensively so this level of segregation will allow for multiple reef-pi devices data analytics.

 

[InactiveAcct]

Wow… I didn’t realize raspberry pi’s were so scarce with the chip shortage. Can’t find them anywhere for less than $100, huge bummer for some projects I had planned

 

[Coolhack]

Simple two questions, what is the best raspberry pi camera to use with reef pi and also how many cameras can I have hooked up at one time?

 

[buddy.]

I am getting a bug where my temp graphs on my dashboard seem to not be updating. I was on 4.3 and upgraded to 5.1 but its still happening.
I know its still monitoring the temp in real time but the graphs are always behind by 20 - 30 min and only seem to randomly update. Perhaps I am missing a setting somewhere?

 

[Ranjib]

Ok I will post an issue to track it. I use Grafana extensively so this level of segregation will allow for multiple reef-pi devices data analytics.

If you use grafana and Prometheus you can do this already, because Prometheus will let you filer by job Id (pi in our case )

 

[Ranjib]

I am getting a bug where my temp graphs on my dashboard seem to not be updating. I was on 4.3 and upgraded to 5.1 but its still happening.
I know its still monitoring the temp in real time but the graphs are always behind by 20 - 30 min and only seem to randomly update. Perhaps I am missing a setting somewhere?

Hmm. I can’t think of anything that will do this ,, if you refresh the graphs get updated ?

 

[buddy.]

Hmm. I can’t think of anything that will do this ,, if you refresh the graphs get updated ?

The only way i can get them to update is to reboot the pi. The graph has not updated for 3 hours at this point.

 

[GaryE]

The only way i can get them to update is to reboot the pi. The graph has not updated for 3 hours at this point.

point a browser at the reef-pi endpoint for stats http://REEFPI/x/metrics change this to the appropriate host name or ip address you use.

Refresh that page a few times and watch the memory numbers :


go_memstats_alloc_bytes 7.326184e+06
# HELP go_memstats_alloc_bytes_total Total number of bytes allocated, even if freed.
# TYPE go_memstats_alloc_bytes_total counter
go_memstats_alloc_bytes_total 9.1560913656e+10
# HELP go_memstats_buck_hash_sys_bytes Number of bytes used by the profiling bucket hash table.
# TYPE go_memstats_buck_hash_sys_bytes gauge
go_memstats_buck_hash_sys_bytes 938885

They should change a bit every refresh.. If they do, reef-pi is likely not where you are having the issue.

 

[Ranjib]

point a browser at the reef-pi endpoint for stats http://REEFPI/x/metrics change this to the appropriate host name or ip address you use.

Refresh that page a few times and watch the memory numbers :


go_memstats_alloc_bytes 7.326184e+06
# HELP go_memstats_alloc_bytes_total Total number of bytes allocated, even if freed.
# TYPE go_memstats_alloc_bytes_total counter
go_memstats_alloc_bytes_total 9.1560913656e+10
# HELP go_memstats_buck_hash_sys_bytes Number of bytes used by the profiling bucket hash table.
# TYPE go_memstats_buck_hash_sys_bytes gauge
go_memstats_buck_hash_sys_bytes 938885

They should change a bit every refresh.. If they do, reef-pi is likely not where you are having the issue.

This is reef-pi dashboard related or grafana/prom related ?

 

[GaryE]

This is reef-pi dashboard related or grafana/prom related ?

I'd say if his metrics endpoint is emitting data every refresh, it'd be an issue with either prometheus or grafana..

 

[Ranjib]

I'd say if his metrics endpoint is emitting data every refresh, it'd be an issue with either prometheus or grafana..

The dashboard is independent of that . When browser is open, the charts should auto update every 10s. The metrics itself gets updated based on the individual module’s interval time (like ato interval )

 

[GaryE]

The dashboard is independent of that . When browser is open, the charts should auto update every 10s. The metrics itself gets updated based on the individual module’s interval time (like ato interval )

Hmm.. I was stuck on Grafana.. I guess he's talking about the reef-pi graphs.. Ooops.. ;)

 

[marekd1]

If you use grafana and Prometheus you can do this already, because Prometheus will let you filer by job Id (pi in our case )

In my case I use NodeRed and all DB functions go to InfluxDB. Grafana pull data from InfluxDB. The NodeRed also does bunch of logic functions and is my central IoT processing engine. Having this function native at the device level will minimize the external app sprawl and management nightmare. I host all of these as containers on Docker locally. This is the only reason I've asked if this could be done at the Reef-Pi level.

 

[GaryE]

Working on a display for the outside of my project case so I can get a quick look at some parameters.

Currently running on a raspberry pi with python. Going to convert to Arduino...

 

[Ranjib]

In my case I use NodeRed and all DB functions go to InfluxDB. Grafana pull data from InfluxDB. The NodeRed also does bunch of logic functions and is my central IoT processing engine. Having this function native at the device level will minimize the external app sprawl and management nightmare. I host all of these as containers on Docker locally. This is the only reason I've asked if this could be done at the Reef-Pi level.

You are not using any less software right ? Instead of influx, we use Prometheus (for long term retention and consolidation of metrics ) .

I thought about this before, would it be possible to make what is displayed dependant on the time-width of the graph? I.e. only display hours and minutes OR day and month OR just month OR month and year OR just year depending on how much time the graph covers. The full timestamp can still be left in the popup whn selecting datapoints to clear up confusions

Or something similar, to declutter irrelevant data

we can. Because it’s software , we can definitely do something clever to detect the available space and adjust accordingly . But we have to test it , and be mindful of responsive scenarios (phone and tablets )

 

[Ranjib]

In my case I use NodeRed and all DB functions go to InfluxDB. Grafana pull data from InfluxDB. The NodeRed also does bunch of logic functions and is my central IoT processing engine. Having this function native at the device level will minimize the external app sprawl and management nightmare. I host all of these as containers on Docker locally. This is the only reason I've asked if this could be done at the Reef-Pi level.

How you are moving the metric from reef-pi to influxdb ? You can not assign the metadata in that mechanism ?
it’s bit confusing on reef-pi end to specify those metadata since it’s not pushing metrics anywhere, it just publish the metrics in a handler/endpoint which prom scrapes . Any metadata is associated by the prom scrape config