AWS Athena
×
Papercrane
Where does an Athena query go after it runs?
Into the results bucket, as a CSV with a generated name. Papercrane's agent writes the same SQL, partition predicate included, and publishes it as a hosted dashboard that refreshes on a schedule, so the answer outlives the query.
Connect in about a minute with a key you create in your AWS settings. Free to connect and ask. Paid plans start when you publish a dashboard. See pricing
The afterlife of a query
You have run this exact query before. On Athena that is not a rhetorical point: running it again scans the lake again, and every scan is billed.
One query
you have run before
the path it takes today
1
Run in the console query editor
2
s3://aws-athena-query-results-…/Unsaved/2026/08/f3a9c1e2….csv
3
Download results → pasted into a sheet
4
A screenshot of the sheet, in a slide, in Slack
5
Stale by Thursday. Someone asks again. The lake is scanned again, at $5 a TB.
the same query, published
1
The agent writes it, partition predicate in the WHERE
2
It becomes a tile on a hosted dashboard, refreshed on the schedule you set
3
Everyone reads the same live page. One scan per refresh, whoever is looking.
Athena prices every scan, so the most expensive habit on a lake is answering the same question twice. A published dashboard scans once per refresh and serves everyone from the results; set the schedule to the cadence of the question, since an hourly refresh nobody reads costs more than the habit it replaced.
A ceiling AWS enforces, not us
Give Papercrane its own workgroup and set a per query scan limit; Athena cancels anything that crosses it, in your account, on AWS's side of the line. The agent writes the partition into the filter, and every query's SQL is yours to read before you trust it.
Ask what the lake already knows
Error rate by service from the ALB logs, this week against last, one chart.
Product events in the lake joined to Stripe revenue: which features do paying accounts actually use?
The same dashboard against your dev and prod catalogs, side by side.
Every answer arrives with its SQL attached, and the dashboard it lands on is code you own.
How it works
Connect with a key you create in your AWS settings.
About a minute. Pick the workgroup and catalog it runs in; connect more than one and address each by name.
The agent reads your Glue catalog.
Databases, tables, and partition keys, then SQL you can read, with joins across databases when the answer needs them.
Publish a hosted dashboard on a refresh schedule you set.
Share it by link, embed it, or put it on a custom domain. When a partition lands with a drifted schema and a query dies, the error routes back to the agent, and the repair is a change to code you can read.
Your data stays in S3. Dashboards run their SQL at refresh time and serve the results to viewers.
When you don't need us
If everyone who needs the dashboard has an AWS login, QuickSight readers are $3 a month, live inside your account, and speak IAM natively: cheap internal readers are its honest win, and the clunky authoring may be worth it at that price. If what you watch is operational time series and your on call already lives in Grafana, the Athena plugin is free and puts lake queries next to your alerts. Papercrane earns its keep when the answer joins sources outside the lake, or goes to someone who will never have an AWS login.
Questions Athena users ask us
The lake holds the events. The money and the traffic usually live somewhere else.
Connect the lake, ask the question you answered last week, and this time publish it.
Free to connect and ask. Paid plans start at publishing. See pricing
Papercrane
Integrations: