// the find
grafana/grafana-infinity-datasource
API datasource for grafana. Visualize data from JSON / CSV / TSV / XML / GraphQL endpoints
A Grafana datasource plugin that pulls data from arbitrary HTTP endpoints (JSON, CSV/TSV, XML, GraphQL, HTML, RSS/Atom) and turns it into Grafana data frames, with auth for everything from API keys to AWS, Azure and GCP signing. It is for people who want to chart an internal API or a SaaS endpoint without writing their own plugin. The repo was pushed yesterday and sits under the grafana org, so it is actively maintained.
Queries run in the Go backend (pkg/pluginhost, pkg/infinity), not the browser, and the HTTP layer lives in pkg/httpclient with its own OAuth2 and TLS tests, so credential handling is in tested server code. The golden-file suite under pkg/testsuite/golden/parsers crosses every format with every query mode in both inline and URL form, so a parser regression shows up as a diff rather than a broken dashboard. Auth coverage is wide for one plugin, from basic and API key through OAuth2 JWT, digest, and AWS, Azure and GCP signing, which means an endpoint behind any of them is configuration rather than new code.
Each format comes in five modes (backend, simple, UQL, GROQ, jq-backend), and the README adds JSONata to the list without saying which to pick. A new user has to choose a query dialect before they know what they need, and every new format has to be tested against all five. The README is a feature list that links out; it has no example query, config snippet, or sample output, so adopting this means learning its behavior from grafana.com rather than from the repo. HTML scraping is the weakest format on the list, since selector-based extraction tends to break when page markup changes, and nothing in the README warns about that for anyone planning a dashboard that has to run unattended.