Lost Gateway connection field in datasets tab

[bsztaje]

I have no idea what I did wrong (and probably I mixed up a lot) but the  Gateway connection field from the datasets tab in the service disappeared and I don't know how to fix it.

Of course it was there and worked fine before.

I removed everything and reinstalled it but it didn't help.

I use a personal gateway and it works fine locally. It is also visible on service.

Service:

has anyone come across this?

Thanks in advance!

[ibarrau]

Hi. The error is not about the gateway. The problem is that PowerBi Service can't recognize the data source. Because it wan't understand the source it won't show the gateway. This is a common error for APIs or web scrapping.

Check your source, it says it's a dynamic one. Feel free to come back with more insights about your sources (which one is it or power query first step)

I hope that helps,

[bsztaje]

Hi,

Thanks for answer, but I'm not sure what you mean by source checking.

The data source is python script, and reading to power BI works locally on the desktop.

All privacy levels set to public.

[ibarrau]

Ok, you are using python to get data. Now. What data? Power Bi can't identify that data. Let's see if we can connect in a different way first and then you can use python for transformations if you need.

Regards,

[bsztaje]

I'm reading data from JIRA using script below.

(sensetive data removed 😉 )

as I said - both reading and refreshing with this script works locally.

Earlier also works fine in service- same script, same data.
I don't know what I have changed.

import pandas as pd
from jira import JIRA

jira = JIRA(options={'server': 'xxx'}, basic_auth=('xxx', 'xxx'))

data = []
for singleIssue in jira.search_issues(jql_str='issuetype = Story AND Team = 1340 AND (status ="Ready for Assignment" or (status = Backlog and substatus = ready )) AND "Acceptance Criteria" IS NOT EMPTY AND description IS NOT EMPTY and "epic link" IS NOT EMPTY AND "Story Points" !=0 AND ("Story Flags" !=Planned OR Labels !=planned) ORDER by Rank DESC'😞
    data.append([singleIssue.key, singleIssue.fields.summary, singleIssue.fields.status, singleIssue.fields.customfield_10023, singleIssue.fields.customfield_10001, singleIssue.fields.customfield_10014])

StoryReady2Assigment = pd.DataFrame(data, columns=['key', 'summary','status', 'story points', 'team', 'EpicLink'])

data = []
for singleIssue in jira.search_issues(jql_str='issuetype = Story AND Team = 1340 AND status NOT IN (Closed, Done)'😞
    data.append([singleIssue.key, singleIssue.fields.summary, singleIssue.fields.status, singleIssue.fields.customfield_10023, singleIssue.fields.customfield_10001, singleIssue.fields.customfield_10014])


StoryNotDoneClosed = pd.DataFrame(data, columns=['key', 'summary','status', 'story points', 'team', 'EpicLink'])

[ibarrau]

Are all your tables like that? It looks good. Please double check that ALL the tables in the model are with public privacy level.
I'll ask around because it's looking good. It's not the best idea to use python at power bi to get data from an API but it should work.

As an alternative I can tell you that I was able to connect to Jira with just Power Query and the API. Example:

= Json.Document(Web.Contents("https://[CompanyName].atlassian.net", [RelativePath = "rest/agile/1.0/board"]))

 You just need a secret key that can be generated from your profile.

I hope that helps,

[bsztaje]

Hi, 

You gave me the idea to try with a different data source.

So I prepared the simplest excel file with a simple table as a data source and loaded it to the service as a new .pbix file (no python data)
It worked and Gateway connection appeared.

Then I added the data from the python source to the same file, and it still works!

One step forward 😄

Thanks!

But.... when I want to update source credentials I get this:

how to provide the installation path in the dialog?

There is no possibility of that!

Now left only one python installation (3.8 native / no venv) on the machine where the personal gateway is installed. (previously I had 3 python instances (2.7, 3.8 and 3.10) 

[ibarrau]

Interesting. Be sure that you can open the pbix file from the VM with the gateway and refresh the data. Configure the Python path inside the pbix file settings on the VM. Then publish from there. I guess that way it might read it.
If it doesn't take it from there it might use the regular python (not things like conda) installation on a C folder. Be sure that you have pandas and jira in there too.

I hope that helps,

[bsztaje]

I tried all of the above - vm as well as regular python installations on c: \ python (i only use regular) and nothing. Zero impact. I give up