Directly import a .json file from your workstation. No external authentication required.
For the general flow of adding a source, see Connect Data Sources.
Before you start
A valid .json file: an array of objects
UTF-8 encoding
Step 1 — Prepare your .json file
The simplest shape is an array of objects at the root. A single object is accepted too and gives one row:
[
{"name": "Alice", "address": {"city": "Paris"}, "roles": ["admin"]}
]Nested objects become composed columns (address_city). Arrays are kept as they are in a jsonb column, queryable with ->, ->> and jsonb_array_elements.
Envelopes are read through: {"data": [...]} imports as it is, and when a file holds several collections — a Zabbix export and its hosts, groups, templates — they are listed with their record counts, and every collection you tick becomes its own table.
NDJSON (one object per line) is accepted too.
Step 2 — Upload in HARVEN
Open Workspace → Data sources → Add → JSON
Enter a name for the connector
Drag and drop or select your .json file
Review the detected columns preview
Submit: the rows are loaded into the table
What to enter in HARVEN
JSON file — File to import, UTF-8, array of objects (your workstation)
What HARVEN reads
(dynamic) — Every key found in the file is imported, including keys absent from the first element
For a flat, tabular export, CSV Import is the simpler route.
Need help? Contact [email protected].