Per tenant or bucket network metrics

[cursive siren]

We have just started using the StorageGrid, and we're and MSP. So of course different teams and custoemr want's info how their part is doing.
Is there any way, or is there something in the road map to be able to use the metrics to see; req per second, MB/s, and thouse types of metrcs either per tenant or even per bucket?
Also would be nice so that this can be integrated with Harvest and the NAbox

[wet pecan]

Harvest can pull SG metrics. You can also just make the internal prometheus in SG available. You'll have to make your own per customer dashboards probably.

[cursive siren]

Harvest can, but not the data we're looking for:
Talkted to someone in the havest channel as well and result is that it can't scrape the necessary data, yet at least
https://discord.com/channels/855068651522490400/1354428464664674575

[wet pecan]

i was just thinking storage usage. Per bucket network traffic would be pretty hard to graph outside of loadbalancer endpoints and even there it doesn't seem like it would be terribly easy

iirc, I think someone wrote a log parser that could get you close

might be in the channel history here even

[edgy orchid]

Havest is just Prometheus and grafana. SGWS has its own Prometheus scrappable data. information on that is here https://docs.netapp.com/us-en/storagegrid-enable/tools-apps-guides/federate-prometheus.html#create-a-grafana-dashboard-for-storagegrid

and metrics you can get from it are here https://docs.netapp.com/us-en/storagegrid-116/monitor/commonly-used-prometheus-metrics.html

though I don't see what you are looking for and am not sure how you would break out just that information.

sorry.. https://docs.netapp.com/us-en/storagegrid/monitor/commonly-used-prometheus-metrics.html there that is current

[honest cypress]

I used a combination of restapi, shell and chart.js to create lots of different views for storagegrid, once you get your head around the metric collection its fairly simple.

[sacred sorrel]

Hi @honest cypress, do you have any views that would be generally useful and that you are willing to share?

[honest cypress]

Hi Chris, yes sure, were you after anything in particular? I am still looking for a new role so I have a bit of time on my hands but no curent access to a grid. I can search out my cgi scripts though.

[sacred sorrel]

Hi dids, Nothing in particular. I work on Harvest and we're always looking for useful SG dashboards. I'm not sure if you've seen the dashboards that Harvest already ships with, but we have an issue were folks comment on SG dashboards for Harvest https://github.com/NetApp/harvest/issues/1420

[honest cypress]

Hi Chris, ah OK, i understand. Mine was a more bespoke solution for our operations team, web front end, shell/cgi/restapi into a single pane view for around 7 grids globally. The netapp team at the time were snowed under with requests for tailormade reports etc and my manager asked me to have ago at putting something together. It included graphs/charts for node util space and performance, customer bucket/object usage for a parameter provided number of days history, this was useful for billing queries and delete queue issues in old versions of SG OS. Capacity planning really found it useful as it was something they had previously for ontap but not SG. Will take alook at the link you sent, thanks.

[sacred sorrel]

thanks, that makes sense. Here are a few screenshots from Harvest's three SG dashboards.

[honest cypress]

looks great, unfortunately i dont have a big enough machine at home to run the virtual instance of grid alongside my ontap simulators.

[sacred sorrel]

right!

[fiery fractal]

Hello,
we are looking more visibility of S3 metrics from de SGRID. Currently, we can't see metrics of number operations, type, code of reponse. Our idea with this information is create new dashboards with filter as bucket/tenant/action and with the Top operations, tables, graphs, etc. From REST-API I couldn't find anything related with my question, harvest team can't implemented anything since information doesn't exist in the internal prometheus and even in the REST-API for consul it. SGRID team has plans to implement this information to consul it from any way?