I have a customer who is choosing between having “User-based pricing” versus “Capacity-based pricing” for their reader users. I did a lot of research about pricing, reader sessions, etc, but got confused at the end and would like to clarify a few questions related to both pricing and metrics:
For “User-based pricing” the price is fixed $3 per user per month regardless of user being active or inactive, is it correct?
For “User-based pricing” we don’t have any concept of a reader session, it’s $3 per month no matter how long the user interacted with dashboard(s), correct?
For “Capacity-based procing” we have a price per number of sessions (min 500sessions/$250/month). Can I clarify here are 500 sessions meant for a single user, or just 500 sessions across all users, meaning that basically per session price is $0.5)?
For “Capacity-based procing” - what if 500 sessions were purchased but as a result more sessions are needed, does this mean that any extra sessions will go for $0.5?
Is there any max $number per user for “Capacity-based procing” (e.g. how it was before for a user-priced approach $0.3 per session and up to max $5 regardless how many sessions reader user had?)
For “Capacity-based procing” - imagine user opened a few dashboards in each tab. The moment a user opens a dashboard 30 mins session starts, correct? So, for 5 dashboards is it considered as one session or we consider session is linked to a one dashboard only, meaning that we have 5 sessions for a user?
For “Capacity-based procing” - even if the user session lasted less than 30 mins (e.g. just 5 mins), the cost will be still as for 30 mins session, correct?
And also I have a question about QS session metrics. I read from different places that such metrics don’t exist, correct? We need a metric to show total number of sessions (but if I’m not mistaken, this somehow could be taken from CUR, correct?) but also we need to have a metric user/#of sessions, and this I think can’t be derived from anywhere? I know that from CloudTrail we have a GetDashboard event, which means the user opened a dashboard, but it’s not related to the session.
I’ve actually had a look to the CUR and also saw this QuickSight now provides standardized user level cost and usage data - so it seems like now from CUR we can actually calculate the number of sessions per user using line_item_usage_type = QS-Reader-Usage-Bonus-Session and calculating the SUM(line_item_usage_amount).
Correct me if I’m wrong please!
And also, in addition to my questions above, can you please explain what QS-Reader-Usage-Bonus-Session and QS-Reader-Usage-Cap-Session mean?
You’ve asked some great questions about QuickSight pricing and metrics. Let me address each of your points:
User-based pricing: Correct, it’s $3 per user per month, regardless of activity.
User-based pricing and reader sessions: Correct, there’s no concept of reader sessions. It’s a flat $3 per month per user.
Capacity-based pricing: The 500 sessions are across all users, not per user. So yes, it’s essentially $0.5 per session. Here is addtional information on capacity pricing breakdown.
Exceeding purchased capacity: Yes, additional sessions beyond the purchased capacity are charged at $0.5 per session.
Maximum cost per user in capacity-based pricing: There’s no maximum cost per user in capacity-based pricing. The $5 cap was for the older reader session pricing model.
Multiple dashboards and sessions: My understanding of the definition of a user session in Quicksight is any time an individual interacts with a dashboard regardless if its in the Quicksight UI or embedded. A Reader Session starts with user-initiated action (e.g., login, dashboard load, page refresh, drill-down or filtering) and runs for next 30-minutes. Keeping Amazon QuickSight open in a background browser window/tab does not result in active sessions until the Reader initiates action on page.
Session duration: Correct, even if a session lasts less than 30 minutes, it’s still counted (and charged) as a full session.
Regarding QuickSight session metrics:
please visit these two links below
The only concern is about question 6 - I got an information that it seems like during 30 mins user can open as many dashboards as he wants and it will be considered as 1 session, however user can’t work on all the tabs highly likely simultaneously, meaning that some initially opened dashboards will be hanging in a background and new session will start for them as soon as the user returns back to work with the dashboard.
Also, the first metric I think is more about the number of logins, not the number of reader sessions, but the second one is what’s needed!