
Tuomas Savonius
Joined Apr 15, 2021
·
Last activity Dec 31, 2024
Following
0
Followers
0
Total activity
45
Votes
3
Subscriptions
18
ACTIVITY OVERVIEW
BADGES
ARTICLES
POSTS
COMMUNITY COMMENTS
ARTICLE COMMENTS
ACTIVITY OVERVIEW
Latest activity by Tuomas Savonius
Tuomas Savonius commented,
I tried using secure settings too, will not work because:

View comment · Posted Dec 31, 2024 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius commented,
Any updates? Why is this blocked in Zafclient?
View comment · Posted Dec 23, 2024 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius commented,
The issue is still present as of December 2024. I encountered the same "AttachmentUnprocessable" error (status 422) when trying to upload files using the ZAF client. I implemented the file upload using the regular uploads endpoint like this:
const uploadRequest = {
url: `/api/v2/uploads?filename=${file.name}.${file.fileExtension}`,
method: 'POST',
contentType: 'application/binary',
data: file.url // ArrayBuffer from FileReader
};
Even with proper ArrayBuffer handling through FileReader and following the API documentation exactly, the ZAF client still fails to process the attachment. This confirms what @Shawn mentioned earlier about ZAF client not supporting binary file uploads, and the issue appears to remain unfixed despite the recent update.
View comment · Posted Dec 23, 2024 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius commented,
I am not a python dev, but I have hit this 1000 hard limit in my C# apps too.
Here's how we cleverly bypass this restriction:
-
Time Windowing: Instead of trying to fetch all tickets at once, we use a moving time window.
- Start with the current date and time as the end of our window.
- Set a start date some time in the past (e.g., 30 days ago).
- Fetch tickets within this time window.
-
Pagination Within Windows: For each time window, we use pagination to get all results.
- This ensures we get all tickets within a specific time range.
-
Moving Backwards: If we haven't reached our desired number of tickets:
- Move the time window back (i.e., set the end date to our previous start date).
- Repeat the process with the new time window.
-
API Limit Detection: We check for the specific API limit message in responses.
- If hit, we adjust our time window and continue.
-
Deduplication: We keep track of seen ticket IDs to avoid duplicates.
- This is crucial as time windows may overlap.
-
Sorting: We sort results by 'updated_at' in descending order.
- This ensures we always get the most recently updated tickets first.
- This ensures we always get the most recently updated tickets first.
I tried to make a python script with this approach, please keep in mind the code is untested and I am not a python expert!
import requests
from datetime import datetime, timedelta
from typing import List, Dict, Any, Tuple
import logging
import sys
import time
class ZendeskTicketFetcher:
def __init__(self, subdomain: str, email: str, api_token: str):
self.base_url = f"https://{subdomain}.zendesk.com/api/v2"
self.auth = (f"{email}/token", api_token)
self.logger = logging.getLogger(__name__)
def get_solved_tickets(self, max_tickets: int = 2000) -> List[Dict[str, Any]]:
end_date = datetime.utcnow()
start_date = end_date - timedelta(days=30) # Start with a 30-day window
results = []
seen_ticket_ids = set()
while len(results) < max_tickets and start_date > datetime(2000, 1, 1): # Arbitrary lower bound
tickets, next_page, api_limit_hit = self.fetch_tickets(start_date, end_date)
if api_limit_hit:
end_date = start_date
start_date -= timedelta(days=30)
continue
new_tickets = [t for t in tickets if t['id'] not in seen_ticket_ids]
results.extend(new_tickets)
seen_ticket_ids.update(t['id'] for t in new_tickets)
while next_page and len(results) < max_tickets:
tickets, next_page, api_limit_hit = self.fetch_tickets(start_date, end_date, next_page)
if api_limit_hit:
break
new_tickets = [t for t in tickets if t['id'] not in seen_ticket_ids]
results.extend(new_tickets)
seen_ticket_ids.update(t['id'] for t in new_tickets)
if len(results) < max_tickets:
end_date = start_date
start_date -= timedelta(days=30)
return results[:max_tickets]
def fetch_tickets(self, start_date: datetime, end_date: datetime, page_url: str = None) -> Tuple[List[Dict[str, Any]], str, bool]:
try:
if page_url:
response = requests.get(page_url, auth=self.auth)
else:
url = f"{self.base_url}/search.json"
query = f"type:ticket updated>{start_date.isoformat()} updated<{end_date.isoformat()} status:solved"
params = {
'query': query,
'sort_by': 'updated_at',
'sort_order': 'desc'
}
response = requests.get(url, auth=self.auth, params=params)
response.raise_for_status()
data = response.json()
api_limit_hit = self.check_api_limit(response.status_code, response.text)
if api_limit_hit:
return [], None, True
return data['results'], data.get('next_page'), False
except requests.RequestException as e:
self.logger.error(f"Failed to fetch tickets: {str(e)}")
sys.exit(1)
def check_api_limit(self, status_code: int, response_content: str) -> bool:
if status_code == 200:
return False
if "Requested response size was greater than Search Response Limits" in response_content:
self.logger.warning("API limit hit. Adjusting time window.")
return True
self.logger.error(f"Failed to get ticket data from Zendesk. Status code: {status_code}")
self.logger.error(f"Response content: {response_content}")
time.sleep(2)
sys.exit(1)
# Usage
logging.basicConfig(level=logging.INFO)
fetcher = ZendeskTicketFetcher('your_subdomain', 'your_email', 'your_api_token')
solved_tickets = fetcher.get_solved_tickets(2000)
print(f"Retrieved {len(solved_tickets)} solved tickets")
View comment · Posted Jul 17, 2024 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius commented,
No updates? I think this is a massive flaw with the ZAF client / secure string implementation.
View comment · Posted Jul 17, 2024 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius commented,
This is a absolute disaster, because using another HTTP client makes the whole process insecure.
https://support.zendesk.com/hc/en-us/community/posts/5200263029914-Uploading-attachments-securely-to-a-ticket
View comment · Posted Jul 17, 2024 · Tuomas Savonius
0
Followers
1
Vote
0
Comments
Tuomas Savonius commented,
This affect us as well, had to make a new view for agents, shame this functionality was so good in the previous dashboard.
View comment · Posted Jan 05, 2024 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius commented,
Do you have any updates on this? This is a high concern item for us currently
View comment · Posted Jul 06, 2023 · Tuomas Savonius
0
Followers
1
Vote
0
Comments
Tuomas Savonius commented,
Duh what a stupid mistake! That fixed the issue. Funny that the adding worked even if the data was given as a string tho!
View comment · Posted Jun 01, 2023 · Tuomas Savonius
0
Followers
0
Votes
0
Comments
Tuomas Savonius created a post,
for some reason, the ticket.collaborators.remove does not work, the adding works fine. No errors in console, verified to happen with latest mozilla, chrome and edge browsers. The id's are 100% correct.
Any idea why this is the case?
installationNeedChanged() {
if (!this.installOrderOptions || this.installOrderOptions.noDevices) {
// THIS DOES NOT WORK (Console logs the removing collaborators, and no erros but the collaborators are not removed)
console.log('removing collaborators');
this._client.invoke("ticket.collaborators.remove", { "id": "367232408557" });
this._client.invoke("ticket.collaborators.remove", { "id": "367232408417" });
} else {
// THIS WORKS correctly
console.log('adding collaborators');
this._client.invoke("ticket.collaborators.add", { "id": "367232408557" });
this._client.invoke("ticket.collaborators.add", { "id": "367232408417" });
}
},
Posted May 31, 2023 · Tuomas Savonius
0
Followers
2
Votes
2
Comments