https://www.umh.app logo
Join Discord
Powered by
Hey guys, long time no talk. I'm working on finall...
# umh-support
c
Hey guys, long time no talk. I'm working on finally adding some SHARCs to one of my beta clients that will use UMH. I have a couple of odd little issues.
Copy code
javascript msg.payload ={

    "timestamp_ms": Date.now(), 
    "count": msg.payload.v.s1.v,
}
msg.topic = "ia/factoryinsight/Aachen/testing/processValue";
return msg;
Posting that in my Nodered....
using that in my Grafana...
No time bucket.
More info hang on....
my queries are getting slow.
This one was fast....
but a few days after running they started getting slow. I'm pushing the count into timescaledb everytime we publish out from the SHARC.
And this is for the week. This took about 10 seconds to return the data for last week from 6-20 to 6-24
and the last 7 days took.....
oh man it's still running.
How can I take a look at the data to get an idea of what is going on?
Usually I would go into PGAdmin or DataGrip and poke around the db to see what the data looks like, indices, time buckets, so on.
j
There is a specific count message feature which should be optimized. Basically, just replace processValue with count
Then you should also be able to leverage the automated calculated KPIs such as productionSpeef
Speed
you can also get somehting like this graph with using the "accumulatedProducts" feature:
although only actual output will output something
i am currently testing it with your 300k count if there are any performance problems
with this it it working quite fast, even with 190 mn counts
to take a look yourself at the database, check out this: https://umh.docs.umh.app/docs/production-guide/administration/access-database/
c
Thanks Jeremy.
j
Can also setup tomorrow a call to help you getting started? Or a small training.
c
I'll be at another client tomorrow installing a SHARC and I'm not sure when I will be onsite yet.
I would love to do that though.
Let me see what I can get in order today.
quick question, does the customerid/location have to exist or will it be created dynamically if it comes in on the topic?
Copy code
ia/<customerID>/<location>/<AssetID>/count
j
factoryinsight should always be factoryinsight (if you want to kee the default data source and not create your own one). location and asset will be created dynamically
might take a couple of seconds
but they will popup automatically
c
What would you expect "this" to do ---
Copy code
"ia/factoryinsight/CCS/count"
As far as selection in grafana
Which datasource should I be using?
I don't get CCS as an option
j
you are missing one additional information about the asset
ia/factoryinsight/whateverlocation/CCS/count
if you are worried about what is happening in the background ,you could check the logs of mqtt-kafka-bridge (when using MQTT) or kafka-to-postgresql
it should output you if a message is not up to standard
c
Copy code
ia/factoryinsight/Wallace/CCS/count"
which datasource is that going to go into? umh-datasource or umh-v2-datasource?
j
both datasources, but i would use v2
more modern and the defau;t opne
(both goes into timescaledb in the background, its jsut two version on how to access our API between grafana and timescaledb)
c
@Jermuk I'll be on-site at this customer tomorrow (6/29). Let me know if you have time to help me out. We can setup a Google Meet or something.
j
I am available today at for an hour, and then only again for 30min after
but I assume @davidmackel or @krgrsebastian can help you too in setting up the "standard" stuff with counts, states, etc. in Node-RED
c
I really need some help. I'm noticing broker connection issues from nodered and my topic data is not making it's way into grafana.
I'll be on-site for another 3 hours.
The top one is umh-datasource and the bottom one is umh-v2-datasource
Their both looking at the same data point.
Copy code
javascript
 msg.payload ={

    "timestamp_ms": Date.now(), 
    "count": msg.payload.v.s1.v,
}
msg.topic = "ia/factoryinsight/Wallace/CCS/count";
return msg;
How do I only use the count and timestamp in the panel? We won't be recording scrap for a long time if ever.
j
I'll be home in like 10 minutes, can do a quick call then?
@chriseudy
10 min will be enough to do a good dashboard
Just prepare your laptop to have everything open: nodered, grafana, and lens
If you give me then 5 more minutes, I'll create a simple oee dashboard
c
Sure
j
2min
c
Now the data is not updating in grafana, but I seee the data going through nodered.
Talk to you tomorrow
j
ok prepared some flows for you and tried them out with a SHARC simulator. could be a good blog article as well to show the dashboards you can now create with the SHARC
i think the best way to go fo is if you could switch the SHARC into accumulation mode
also started a small blog article, where we can write it down and advertise a little bit the SHARC 🙂 https://learn.umh.app/p/7009ea9e-5f28-4b75-ae81-98e27e3c9828/
c
2 Views
PreviousNext
Powered by