[➕ Feature]: Query Jira issues by JQL from both Jira providers

[purisev]

Is your feature request related to a problem? Please describe.

_query in both Jira providers takes a ticket_id and a board_id
(keep/providers/jira_provider/jira_provider.py:652,
keep/providers/jiraonprem_provider/jiraonprem_provider.py:593). With a ticket
id it returns that one issue. Without one it fetches a board's issues and returns
the count alone:

issues = response.json()
return {"number_of_issues": issues["total"]}

The issues it just fetched are thrown away, so a workflow has nothing to iterate,
and the only question it can ask is how many issues a board holds. "Open in
project SA for more than three days" or "assigned to nobody" is a JQL query, and
there is no parameter to send one through.

Describe the solution you'd like

Three parameters on _query: jql, max_results, fields. The result carries
the total, the query it ran, and the issues, so a workflow can foreach over
them:

with:
  jql: "project = SA AND status = Open AND created <= -3d"
  max_results: 50
  fields: summary,status,created,assignee

max_results: 0, the default, answers with the count and an empty list, which
covers a threshold check without moving issue data around. fields keeps custom
fields out of the payload.

The endpoints differ between deployments. Server and Data Center answer both
halves in one /rest/api/2/search request. On Cloud that endpoint is going away,
so the search goes to search/jql, which returns a page and no total, and the
count comes from search/approximate-count beside it.

Describe alternatives you've considered

Fetching a board and filtering afterwards. The board path returns no issues at
all, only their number, so there is nothing to filter.

Additional context

A PR implementing this follows.