Skip to main content

Data Source Configuration

Step 5: Server config - data sources

The OPAL server serves the base data source configuration for OPAL client. The configuration is structured as directives for the client, each directive specifies what to fetch (url), and where to put it in OPA data document hierarchy (destination path).

The data sources configured on the server will be fetched by the client every time it decides it needs to fetch the entire data configuration (e.g: when the client first loads, after a period of disconnection from the server, etc). This configuration must always point to a complete and up-to-date representation of the data (not a "delta").

You'll need to configure this env var:

Env Var NameFunction
OPAL_DATA_CONFIG_SOURCESDirectives on how to fetch the data configuration we load into OPA cache when OPAL client starts, and where to put it.

Data sources config schema

The value of the data sources config variable is a json encoding of the ServerDataSourceConfig pydantic model.

Example value

"config": {
"entries": [
"url": "",
"topics": [
"config": {
"headers": {
"Authorization": "Bearer FAKE-SECRET"

Let's break down this example value (check the schema for more options):

Each object in entries (schema: DataSourceEntry) is a directive that tells OPAL client to fetch the data and place it in OPA cache using the Data API.

  • From where to fetch: we tell OPAL client to fetch data from the API (specifically, from the policy-config endpoint).
  • How to fetch (optional): we can direct the client to use a specific configuration when fetching the data, for example here we tell the client to use a specific HTTP Authorization header with a bearer token in order to authenticate to the API.
  • Where to place the data in OPA cache: although not specified, this entry uses the default of / which means at the root of OPA document hierarchy. You can specify another path with dst_path (check the schema).
  • Which clients should fetch: the entry would be processed only by clients subscribed to one of the specified topics. (If unspecified, the default policy_data topic would be used).
  • How often to fetch: entry that has the periodic_update_interval value would be periodically pushed to clients with that interval (in secs), so they can keep refetching the most updated data from specified source.

Encoding this value in an environment variable:

You can use the python method of json.dumps() to get a one line string:

❯ ipython

In [1]: x = {
...: "config": {
...: "entries": [
...: ... # removed for brevity
...: ]
...: }
...: }

In [2]: import json
In [3]: json.dumps(x)
Out[3]: '{"config": {"entries": [{"url": "", "topics": ["policy_data"], "config": {"headers": {"Authorization": "Bearer FAKE-SECRET"}}}]}}'

Placing this value in an env var:

export OPAL_DATA_CONFIG_SOURCES='{"config": {"entries": [{"url": "", "topics": ["policy_data"], "config": {"headers": {"Authorization": "Bearer FAKE-SECRET"}}}]}}'

Please be advised, this will not work so great in docker-compose. Docker compose does not know how to deal with env vars that contain spaces, and it treats single quotes (i.e: '') as part of the value. But with docker run you should be fine.

Delegating the sources config to an API server

For more dynamic cases where you'd need different clients to load different sets of data at different times; you can have the OPAL-clients be redirected to another server to fetch the configuration. See the details in this tutorial


Since OPAL_DATA_CONFIG_SOURCES often contains secrets, in production you should place it in a secrets store.