Improve validation error messages for DBRP API

[Hualin-AUG]

Steps to reproduce:
List the minimal actions needed to reproduce the behavior.

1. Having an InfluxDB Cloud account (free tier)
2. Setup a bucket and write data into it
3. Try to do DBRP mapping with following command:

curl --request POST 'https://us-west-2-1.aws.cloud2.influxdata.com/api/v2/dbrps'
--header 'Authorization: Token {my token}'
--header 'Content-type: application/json'
--data '{
"bucketID": "{my bucketID}",
"database": "{some database name}",
"default": true,
"orgID": "{my orgID}",
"retention_policy": "30days"
}'
Expected behavior:
A DBRP should be setup for the bucket I give

Actual behavior:
400 Bad request error happens with following message
{
"code": "invalid",
"message": "invalid ID"
}

Environment info:

• InfluxDB version: InfluxDB Cloud

Config:
Token I use an all permission token

[rbetts]

Let's also make the "Invalid ID" error messages disambiguate which ID(s) are invalid. (See re-used errors here: https://github.com/influxdata/influxdb/blob/master/dbrp/http_server_dbrp_test.go#L185).

[timhallinflux]

@Hualin-AUG -- you might also consider trying the updated CLI which now has DBRP commands which should make this a bit easier to manage the mappings themselves. https://docs.influxdata.com/influxdb/v2.0/reference/cli/influx/v1/dbrp/

We'll improve the error messages here in parallel.

[danxmoran]

+1 to more specific error messages. In that same direction, I suspect the actual error here is/was that the Cloud implementation expectsbucket_id and organization_id instead of the Swagger-specified bucketID and orgID (it's fixed in OSS but I'm not sure if it's gone through to Cloud yet).

I don't think either impl is validating that requests contain all the expected keys. When there's a mismatch, the system ends up with an empty string for one (or both) IDs, and the empty string fails validation. The most specific validation error would be something like Missing required key: "xxx"

[danxmoran]

@Hualin-AUG if the new CLI tools don't work for you, could you also try using bucket_id instead of bucketID and organization_id instead of orgID in your request payload?

[Hualin-AUG]

@Hualin-AUG if the new CLI tools don't work for you, could you also try using bucket_id instead of bucketID and organization_id instead of orgID in your request payload?

That absolutely solves the problem.

[rbetts]

@Hualin-AUG thanks for the bug-report. Really helpful as we iron out some usability gotchas here!