pagination

Pagination Skill

This skill covers pagination strategies for the Jira Cloud REST API, including offset-based and cursor-based pagination.

Offset-Based Pagination

Most Jira Cloud REST API endpoints use offset-based pagination with startAt and maxResults parameters.

Parameters

Parameter	Description	Default
startAt	The index of the first item to return (0-based)	0
maxResults	The maximum number of items to return per page	Endpoint-specific (usually 50)

Response Fields

Field	Description
startAt	The index of the first item returned
maxResults	The maximum number of items that could be returned
total	The total number of items matching the query
values / issues	Array of items for the current page

Example: Paginate Through All Issues

# First page
curl -X POST \
  -H "Authorization: Basic $(echo -n '$JIRA_USER_EMAIL:$JIRA_API_TOKEN' | base64)" \
  -H "Content-Type: application/json" \
  "https://{domain}.atlassian.net/rest/api/3/search/jql" \
  -d '{
    "jql": "project = PROJ ORDER BY created DESC",
    "startAt": 0,
    "maxResults": 100,
    "fields": ["summary", "status", "assignee"]
  }'

# Response:
# { "startAt": 0, "maxResults": 100, "total": 350, "issues": [...] }

# Second page
curl -X POST \
  -H "Authorization: Basic $(echo -n '$JIRA_USER_EMAIL:$JIRA_API_TOKEN' | base64)" \
  -H "Content-Type: application/json" \
  "https://{domain}.atlassian.net/rest/api/3/search/jql" \
  -d '{
    "jql": "project = PROJ ORDER BY created DESC",
    "startAt": 100,
    "maxResults": 100,
    "fields": ["summary", "status", "assignee"]
  }'

# Continue until startAt + maxResults >= total

Pagination Loop Pattern

import requests
import math

base_url = "https://your-domain.atlassian.net"
auth = ("[email protected]", "api-token")

start_at = 0
max_results = 100
all_issues = []

while True:
    response = requests.post(
        f"{base_url}/rest/api/3/search/jql",
        auth=auth,
        json={
            "jql": "project = PROJ ORDER BY created DESC",
            "startAt": start_at,
            "maxResults": max_results,
            "fields": ["summary", "status"]
        }
    )
    data = response.json()
    all_issues.extend(data["issues"])

    if start_at + max_results >= data["total"]:
        break
    start_at += max_results

print(f"Fetched {len(all_issues)} of {data['total']} issues")

Bash Pagination

#!/bin/bash
START_AT=0
MAX_RESULTS=100
TOTAL=1  # Will be updated after first request

while [ $START_AT -lt $TOTAL ]; do
  RESPONSE=$(curl -s -X POST \
    -H "Authorization: Basic $(echo -n "$JIRA_USER_EMAIL:$JIRA_API_TOKEN" | base64)" \
    -H "Content-Type: application/json" \
    "$JIRA_BASE_URL/rest/api/3/search/jql" \
    -d "{
      \"jql\": \"project = PROJ\",
      \"startAt\": $START_AT,
      \"maxResults\": $MAX_RESULTS,
      \"fields\": [\"summary\", \"status\"]
    }")

  TOTAL=$(echo "$RESPONSE" | jq '.total')
  COUNT=$(echo "$RESPONSE" | jq '.issues | length')
  echo "Fetched $COUNT issues (startAt: $START_AT, total: $TOTAL)"

  # Process issues here
  echo "$RESPONSE" | jq '.issues[] | {key: .key, summary: .fields.summary}'

  START_AT=$((START_AT + MAX_RESULTS))
done

Server-Enforced Limits

Different endpoints have different maximum page sizes:

Endpoint	Max maxResults
/rest/api/3/search/jql	100
/rest/api/3/user/search	1000
/rest/api/3/group/member	50
/rest/api/3/project/search	100
/rest/agile/1.0/board	100
/rest/agile/1.0/sprint/{id}/issue	100

If you request more than the maximum, the server silently reduces maxResults to its limit.

Cursor-Based Pagination

Some newer endpoints use cursor-based pagination with cursor or after parameters instead of startAt.

Example: Paginate with Cursor

# First page
curl -X GET \
  -H "Authorization: Basic $(echo -n '$JIRA_USER_EMAIL:$JIRA_API_TOKEN' | base64)" \
  "https://{domain}.atlassian.net/rest/api/3/field/search?maxResults=50"

# Response includes a 'nextPage' URL or cursor
# { "values": [...], "isLast": false, "nextPage": "https://...?cursor=abc123" }

# Follow the nextPage URL for subsequent pages

Performance Best Practices

1. Request only needed fields — Use the fields parameter to limit returned data
2. Use reasonable page sizes — 50-100 is optimal; larger sizes increase response time
3. Avoid deep pagination — For large datasets, use JQL with date filters instead of paginating to page 1000+
4. Cache total counts — The total field may be expensive to compute; avoid re-requesting it if unchanged
5. Parallelize cautiously — Concurrent pagination requests may hit rate limits
6. Use POST for search — GET requests with long JQL may exceed URL length limits
7. Add ORDER BY — Without explicit ordering, results may shift between pages if data changes

Handling Data Changes During Pagination

Data may change between paginated requests:

• New items may be added, shifting indices
• Items may be deleted, causing gaps
• Items may be updated, changing sort order

Mitigations:

• Use ORDER BY key ASC for stable ordering by immutable field
• Use created >= "date" filters for time-bounded pagination
• Accept that pagination is eventually consistent, not a point-in-time snapshot