jira.psom_sla
The state of each PSOM ticket’s Jira Service Management SLA clocks, one row per ticket per SLA per snapshot.
Each row is one SLA (sla) on one ticket (key) as it stood on one snapshot date. Every ticket gets a row for every SLA, even if that clock never started. Join to psom on key and date. A row is unique on (key, sla, date).
Keywords: jira, psom, SLA, breach, time to first response, time to done, touchtime, escalation
Provenance
Written by the psom-sync Lambda in the same run as psom, from the ticket’s SLA fields.
- Updated: Daily, at ~06:00 ET
- Retention: 2 years
- Steward: EDCP Data and Research Team
- Contact: Yinlin Zhang (Data and Research team)
Fields
| Field Name | Type | Description |
|---|---|---|
| date | varchar | Partition key: the date the Lambda ran (YYYY-MM-DD). |
| key | varchar | Jira ticket key, e.g. PSOM-42. |
| sla | varchar | Which SLA clock, one of first_response, time_to_done, pso_time_to_done, or escalated_to_pt. See the Notes for definitions. |
| due | varchar | ISO 8601 datetime with UTC offset, e.g. 2026-07-02T09:52:39-0400. For a running clock, the breach deadline. For a clock that has stopped, when it stopped. Null if the clock never started. |
| breached | boolean | Did the clock miss its goal? false covers both a met goal and a clock still inside its goal. Null if the clock never started. |
| goal_minutes | bigint | The goal, in working minutes. It varies between tickets. |
| elapsed_minutes | bigint | Working minutes on the clock as of the fetch. |
| paused | boolean | Is a running clock paused? Null once the clock has stopped. |
| fetched_at | varchar | UTC timestamp when the Lambda wrote this row (YYYY-MM-DD HH:MM:SS). |
Notes
The four SLAs.
first_response: Jira’s “Time to first response”.time_to_done: Jira’s “Time to done”, for the whole ticket. It keeps running while the ticket is escalated and pauses while waiting for the customer.pso_time_to_done: “Time to Done - PSO”, the PSO team’s touchtime. It pauses while the ticket is escalated, or waiting for the customer. It only started on tickets created from about 2026-08-06.escalated_to_pt: “Escalated to PT”, which runs while the ticket is escalated to the authentication team.
Working time. Goals and elapsed time count working minutes on the SLA’s 09:00 to 17:00 calendar, not wall-clock minutes, so a 40-hour goal is one working week.
Goals vary by ticket. The goal on a clock varies by ticket (for time_to_done, 16, 24, 40 and 72 hours all occur). A breach means the ticket missed its own target. To compare speed across tickets, use elapsed_minutes rather than breached.
Which cycle. A Jira SLA can restart when a ticket is reopened. Each row describes the running cycle if there is one, otherwise the last completed cycle. Earlier cycles are not kept.
Backfilled history (through 2026-09-28). Snapshots are daily from 2026-09-29. Before that they were weekly, on Mondays, and were backfilled from the sla_* columns that psom used to carry. The older snapshots only carry first_response and time_to_done, and only due and breached. There are no rows for the other two SLAs, and goal_minutes, elapsed_minutes and paused are null. breached is null in every row of the 2026-06-30 and 2026-07-06 snapshots, which predate a pipeline fix. It is populated from 2026-07-07 on.