fix(rhosak): ent-4615 gibibytes to binary gigabytes

[cdcabrera]

What's included

• fix(rhosak): ent-4615 gibibytes to binary gigabytes

Notes

• Replaces gibibyte with binary gigabytes, and GiB with GB

How to test

Proxy run check

1. update the NPM packages with $ yarn
2. make sure Docker is running, plus on network, then
3. $ yarn start:proxy
4. review the copy updates and tooltips highlighted in the examples

Example

Feb-23-2022.11-33-23.mp4

Updates issue/story

ent-4615

[mirekdlugosz]

This is good to go to stage-beta

[jlprevatt]

2 comments:

FIRST:
For the items with the spelling "Binary Gigabytes," change the spelling to "Binary gigabytes" (assuming that it is a standalone phrase or that it starts a sentence).
Just as we don't capitalize "inches," there's nothing special about the measurement name "binary gigabytes" that makes it capitalized.

SECOND:
Based on something Dan said when reviewing my documentation for RHOSAK, I think there is something that we have to double-check for the RHOSAK instance hours UoM. I don't know if he had the GUI open when reviewing my doc, but either the doc is wrong or the GUI is wrong. I will paste Dan's comment in the main PR.

[jlprevatt]

RE my "SECOND" comment in my review:

When Dan was reviewing the new content for RHOSAK, he sent me the following comment about my text for RHOSAK instance hours (my proposed text first, then his comment):

Instance hours
The subscriptions service measures instance hours of an OpenShift Streams for Apache Kafka service instance by gibibyte hours. A gibibyte hour is a unit of measurement for computational activity, including the availability and activity for processing workloads, generated by that service instance for a total of one hour.

DANGER: I don't think this is correct. This should be a pure unit of time, but could be phrased as "RHOSAK-hours"
SVCMW01882A: Red Hat OpenShift Streams for Apache Kafka (Cluster Fee, hourly)
Note that the description of the charge there is a pure hourly descriptor

So my change for that was to keep the first sentence but end it by just saying "by hours."
It's worth double-checking whether the GUI is wrong or the doc is wrong since the original on-screen instance hours wording (gibibyte hours) made it through wireframing/GUI implementation/QE.

[dozdowski]

RE my "SECOND" comment in my review:

When Dan was reviewing the new content for RHOSAK, he sent me the following comment about my text for RHOSAK instance hours (my proposed text first, then his comment):

Instance hours
The subscriptions service measures instance hours of an OpenShift Streams for Apache Kafka service instance by gibibyte hours. A gibibyte hour is a unit of measurement for computational activity, including the availability and activity for processing workloads, generated by that service instance for a total of one hour.
DANGER: I don't think this is correct. This should be a pure unit of time, but could be phrased as "RHOSAK-hours"
SVCMW01882A: Red Hat OpenShift Streams for Apache Kafka (Cluster Fee, hourly)
Note that the description of the charge there is a pure hourly descriptor

So my change for that was to keep the first sentence but end it by just saying "by hours." It's worth double-checking whether the GUI is wrong or the doc is wrong since the original on-screen instance hours wording (gibibyte hours) made it through wireframing/GUI implementation/QE.

Jamie is correct, and my previous warning remains. The unit "gibibyte" (binary gigabyte, etc) should ONLY attach to (AFAIK) the storage unit of measure. The networking unit of measure is (again AFAIK) actually gigabytes - GB - and the "collective hours of RHOSAK run" unit of measure is just "hours" or a suitable abbreviation. The example here is one RHOSAK instance that runs for 5 hours and one RHOSAK instance that runs for 10 hours results in a "RHOSAK instance hours" count of 15. I'm fine with either instance hours or RHOSAK hours as shorter options.

[jlprevatt]

So instance hours are just plain ol' counting up hours (no fancy "sum over time" math)

Reviewing Dan's recent response in context of what's in the doc...
here
(references to "derived" UoMs = combo units, something like X over time)

data transfer/network IO = counter = gigabytes

• Doc currently says "gibibytes" (NO derived UoM here)
• Needs to be changed to "gigabytes" per Dan's comment
• DOUBLE-CHECKING--not binary gigabytes? So transfer is measured based on 1000 and storage is based on 1024?

data storage = gauge = BINARY gigabyte hours

• Doc currently says "gibibtye hours" (YES derived UoM here, storage uses "sum over time" AFAIK)
• Needs to be changed to "binary gigabyte hours" per Dan's comment

instance hour/RHOSAK hour = on/off switch = hours

• Doc currently says "hours" (NO derived UoM here)
• No doc change

(We can have a conversation whether "RHOSAK-hours" and "RH-whatever" hours are really-super-helpful on a detailed RHM customer invoice (and whether they would be listed like that) or if using that type of name starts to get tedious. I can't see which is true from my lowly perch.)

[jlprevatt]

(I really thought Barnaby said that all Xbyte-based data for this in our tools was gibibytes (and in no place is it recalculated as true gigabytes), so that's why I want to be extra sure on Dan's comment for data transfer above)

[codecov-commenter]

Codecov Report

Merging #895 (a9ea5a9) into ci (66959db) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##               ci     #895   +/-   ##
=======================================
  Coverage   94.06%   94.07%           
=======================================
  Files         126      126           
  Lines        3692     3694    +2     
  Branches     1445     1445           
=======================================
+ Hits         3473     3475    +2     
  Misses        201      201           
  Partials       18       18           

Impacted Files	Coverage Δ
src/config/product.rhosak.js	86.11% <100.00%> (+0.81%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 66959db...a9ea5a9. Read the comment docs.

[cdcabrera]

Ok review updates are here

Transfer

Storage

Instance

Inventory

[cdcabrera]

Since the inventory display has limited column widths, a further tooltip explanation

[jlprevatt]

Final tweaks I see, as we've worked through it in chat:

• So for the transfer section, the "Binary Gigabyte" labels (cards, vertical axis) can be changed to match the capitalization of the others: "Binary gigabytes"
• For the instance hours section, bottom Instance hour tooltip: "Instance hours usage" (plural) to match
• For the inventory section, it's so space constrained we discussed tooltips for the Data transfer and Data storage columns

[cdcabrera]

Transfer cards

Instances tooltip

[jlprevatt]

Reviewed, including most recent screen caps in Conversation section. Approving.

[cdcabrera]

Holding merge into beta env on this until CCS comes up with their documentation release date. We may have to alter where this commit is added either in stable or beta envs since there are commits in beta that are currently API dependent... we'll commence the commit shell game shortly

@mirekdlugosz @ntkathole @jlprevatt @bclarhk @dozdowski